Last updated: November 23, 2022

This post will be the first in a series examining the evolution of Proto-Algonquian (PA) to its daughter languages. We’ll begin simply by covering Proto-Algonquian itself. Little of what is here is new, and in particular the post draws very heavily (as should hopefully be obvious, although I didn’t inline cite everything) from the work of the two most preeminent comparative Algonquianists, Ives Goddard and David Pentland. For those issues that are uncontroversial I’ve tried to give a quick overview; where there’s substantial disagreement on a topic or where my views diverge from the dominant one I’ve provided more extensive commentary. I’ve also provided a fairly long, though by no means exhaustive, list of (morpho)phonological rules operative in PA—both because of their relevance for later sound changes and analogical reshapings, and because I’m not aware of any place where they’ve all been listed together in quite this fashion (though many are discussed in Pentland 1979a, ch. 7, and Pentland 1999:245-255) and in the order in which they must apply.

Table of Contents

The Algonquian Family

The Algonquian language family consists of roughly 35 to 40 languages, depending on how one counts, most of which are now either extinct, moribund, or seriously endangered. The only relatively healthy (though still endangered) languages—with sizeable numbers of speakers and still being learned and actively used by most children—are many varieties of Cree-Innu, some varieties of Ojibwe, and perhaps Kickapoo. The remaining languages range from significantly to critically endangered. There are, however, numerous efforts underway to both pass the surviving languages on to the next generations, as well as to revive a number of the languages which have already become extinct—or, hopefully, merely dormant!

Map of approximate locations of Algonquian languages, plus Wiyot and Yurok, on which see below
Map of approximate locations of Algonquian languages, plus Wiyot and Yurok, on which see below. The map is roughly based on the location of speakers at time of first contact with Europeans, but since the date of contact varied greatly in different areas, it is unavoidably anachronistic in combining locations from multiple different time periods. The expansion and westward movement of certain groups, particularly Ojibwe, Cree, and Cheyenne speakers, is also anachronistic, and they were all further east and more geographically constrained in the 17th and 18th centuries. “MSK” = “Meskwaki-Sauk-Kickapoo (i.e. Meskwakian) and “SNEA” = Southern New England Algonquian; Powhatan is an alternative name for Virginia Algonquian. (Modified from Wikimedia Commons [CC-by-2.0] based on Goddard 1996b; the Wikimedia Commons map is itself based on Goddard 1996b and the maps in Campbell 1997 and Mithun 1999.)

As a matter of convenience, Algonquian languages are traditionally divided into geographical “Plains,” “Central,” and “Eastern” groupings, but of these only Eastern Algonquian is a true genetic subgroup. There are, however, some other (usually low-level) subgroups that can be identified. A vaguely consensus classification of Algonquian languages is:[1]

  • (1) Blackfoot
    • Siksiká
    • Northern Piegan [= N. Piikani = N. Peigan = Aapátohsipikani]
    • Kainaa [= Kainai = Blood]
    • Southern Piegan [= S. Piikani = S. Peigan = Aamsskáápipikani]
  • Core Algonquian
    • Cree-Innu [= Cree-Montagnais-Naskapi = Cree-Innu-Naskapi] (continua of sublanguages)
      • (2) western Cree (not a clade, only defined by not being Eastern Cree-Innu varieties)
        • Plains Cree (Northern Plains Cree, Southern Plains Cree)
        • Woods Cree
        • Western Swampy Cree
        • Eastern Swampy Cree (E. Swampy Cree, Moose Cree)
        • Atikamekw [= Attikamek(w) = Tête-de-Boule Cree]
      • (3) Eastern Cree-Innu
        • East Cree (Northern EC, Southern Coastal EC, Southern Inland EC)
        • Innu-Naskapi [= Montagnais-Naskapi]
          • Innu[-aimun] [= Montagnais] (Old Innu †, Western Innu, Eastern Innu)
          • Naskapi ((Western) Naskapi, Mushuau Innu(-aimun) [= Eastern Naskapi])
    • Arapaho-Gros Ventre (AGV)
      • (4) Arapaho-Besawunena
        • Arapaho (Northern Arapaho, Southern Arapaho)
        • Besawunena †
      • (5) Gros Ventre [= Aaniiih = Atsina = White Clay] †
    • (6) Nawathinehena † (possibly “Arapahoan”)[2]
    • (7)(?) Ha’anahawunena † (possibly “Arapahoan”?)
    • (8) Cheyenne
      • Northern Cheyenne
      • Southern Cheyenne
      • (?) Sutaio †
    • (9) Menominee [= Menomini]
    • Core Central (?)
      • Ojibwe-Potawatomi [= Ojibwean = Ojibwayan]
        • (10) Ojibwe Proper (continuum of sublanguages) [see here for name variants]
          • Northern Ojibwe Proper (Algonquin, Oji-Cree [= Severn Ojibwe])
          • Northwestern Ojibwe
          • North of Superior Ojibwe
          • Saulteaux
          • Border Lakes Ojibwe
          • Southwestern Ojibwe
          • Old Algonquin †, Nipissing [= Nipissing Algonquin]
          • Eastern Ojibwe
          • Old Odawa †, Odawa [= Ottawa]
        • (11) Potawatomi
          • Northern Potawatomi
          • Southern Potawatomi
      • (12) Miami-Illinois [= Myaamia] †
        • Old Illinois (dialectally mixed records)
        • Oklahoma Miami-Illinois (Peoria [= Peewaalia], Miami [= Myaamia], Wea, Piankashaw)
        • Indiana Miami [= Myaamia]
      • Southern Core Central (?)
        • (13) Meskwakian [= Meskwaki-Sauk-Kickapoo = Fox-Sauk-Kickapoo]
          • Meskwaki-Sauk (Meskwaki [= Fox = Mesquakie], Sauk)
          • Kickapoo
          • (?) Mascouten †
        • (14) Shawnee
          • Absentee Shawnee
          • Eastern Shawnee †
          • Loyal [= Cherokee] Shawnee
    • Eastern Algonquian (protolanguage = PEA)
      • Northern New England Algonquian [= Wabanakian] (?)
        • (15) Mi’kmaq [= Micmac = Mi’gmaq = Mi’kmaw = Mi’gmaw]
          • Listuguj [= Restigouche] Mi’kmaq
          • Nova Scotia Mi’kmaq
          • Newfoundland Mi’kmaq †
          • (Other dialects)
        • Abenakian
          • (16) Maliseet-Passamaquoddy
            • Maliseet [= Malecite = Wolastoqey]
            • Passamaquoddy
          • Abenaki [= Abnaki]
            • (17) Eastern Abenaki †
              • Penobscot (Coastal Penobscot, Inland Penobscot)
              • Caniba [= Kennebec River Abenaki]
              • Arosaguntacook
              • Pigwacket
            • (18) Western [= St. Francis] Abenaki †
      • (19)(?) “Etchemin” †
      • Southern New England Algonquian (SNEA)[3] (overlapping dialect continua)
        • (20) Massachusett [= Natick]-Narragansett (dialect continuum) †
          • Mainland Massachusett [= mod. name Wampanoag/Wôpanâak]
          • Island Massachusett [= mod. name Wampanoag/Wôpanâak] (Nantucket Mass., Martha’s Vineyard Mass.)
          • Most of Williams’s “Narragansett” (= Coweset?)
          • Narragansett Proper (= Southern Narragansett)
        • (21) Loup 1 [in “Loup A”] (= Nipmuck?) †
        • (22) Loup 2 [“Loup A”] (= Norwattuck?? Agawam?? Nipmuck??) †
        • (23) Loup 3/4 [“Loup A” + “Loup B”] (= Pocumtuck??) †
        • (24) Loup 5 [“Loup B”] (= Woronoco???) †
        • (25) Mohegan-Pequot †
          • Mohegan
          • Pequot
          • Montauk
          • Stiles’s “Narragansett” (= Eastern Niantic?)
        • (26) Quiripi-Naugatuck [= Wampano] †
          • Quiripi
          • Naugatuck
          • Paugussett
          • Potatuck
        • (27) Unquachog [= Unkechaug] †
        • (?) Shinnecock †
      • (28)(?) Loup 6 [“Loup B”] (= Pojassick???) †
      • Delawaran [= Delaware-Mahican]
        • (29) Mahican [= Mohican] †
          • Eastern [= Stockbridge] Mahican
          • Western [= Moravian] Mahican
        • Delaware [= Lenape]
          • (30) Munsee †
            • Munsee
            • Wampano
            • (?) Wappinger
          • (31) Unami †
            • Unalachtigo
            • Northern Unami
            • Southern Unami [= Oklahoma Delaware]
      • (32) Nanticoke-Conoy [= Maryland Algonquian] †
        • Nanticoke (Nanticoke, Choptank)
        • Conoy [= Piscataway]
      • (33) Virginia Algonquian [= Powhatan] †
      • (34) Carolina Algonquian [= Pamlico = Pampticough] †

If the above list is difficult to follow, you can expand the image below, which presents basically the same tree although without listing dialects, and with a few somewhat uncertain low-level subgroups left unmarked to save space. It also shows the non-genetic, geographically based traditional convenience groupings of “Central Algonquian” and “Plains Algonquian.” For a rotated version of the image, click HERE.

Possible Algonquian family tree
A possible Algonquian family tree. “Plains Algonquian” and “Central Algonquian” are geographical terms of convenience, not genetic subgroups, and in my opinion “Arapahoan” is probably also a term of convenience and not a subgroup. For “Western Algonquian” see below. Click to expand.

As can be seen, Blackfoot is in many ways the most divergent Algonquian language, and there is some evidence that it is Algonquian’s version of Anatolian, with all other Algonquian languages forming a separate clade, “Core Algonquian.” As indicated in the tree, I have also come to believe that all the Core Algonquian languages other than Eastern Algonquian, i.e. “Western Algonquian,” may form a genetic subgroup, or else derive from one a few dialects of Proto-Core Algonquian which shared a set of vowel changes not found in the rest of the family. But several other interpretations are quite possible. For the evidence and much more discussion, see this post.

The Algonquian family is in turn part of the broader Algic family, the other members of which are two languages spoken on the northwest coast of California, Wiyot and Yurok.[4] While Yurok has only a few remaining native speakers, it also has one of the most successful language revival programs among American Indian nations, boasting significant numbers of fluent and semi-fluent second-language speakers. The last native speaker of Wiyot died in the 1960s. Wiyot and Yurok are sometimes grouped together under the term “Ritwan” but it’s still not clear if they are more closely related to each other in a Ritwan subfamily of Algic, or if all three of Wiyot, Yurok, and Algonquian are roughly equally related. If Ritwan is in fact a subgroup, it appears to be a deep one which must have separated not long after Proto-Algic itself broke up. The time depth separating Wiyot, Yurok, and Algonquian is unclear, but the languages are quite divergent, with only something like fifty to a hundred secure cognates known.[5] A few examples of Algic cognate sets:[6]

  • PA *[ne]kot[w]- : Wiyot kuc- : Yurok koht-PAc **kwət- “one”
  • PA *nyi·š[w]- : Wiyot ḍit- : Yurok nVʔ-PAc **nyi:ł- “two”
  • PA *neʔθ[w]- : Wiyot ḍikh- : Yurok nahkṣ-PAc **nəks- “three”
  • PA *nye·w- : Wiyot ḍiyaʔw-PAc **nyæ:ˀw- “four”
  • PA *oθkonih/ liver” : Wiyot ẃətwəḍ “liver” : Yurok [ˀ]wełkwṛn “liver” ← PAc **wə-łkwən-h/ liver”
  • PA *pemyi : Wiyot púʔm : Yurok pemeyPAc **pəˀmy- “grease”
  • PA *-([a·])peθk(-wi)- “stone, metal” : Wiyot pł́ətk “rock,” płəčwi[čač] “pebbles” : Yurok pełko: “pebble” ← PAc **pləłkwi:
  • PA *šek[i]- : Wiyot tik[əl]- : Yurok ʔahk-PAc **łyək- “urinate”
  • PA *watapya : Wiyot táph- : Yurok ˀwohpeyPAc **wə-tăpy- “spruce root”

Broader relationships for Algic with other families such as Salish, Wakashan, Muskogean, or the language isolate Kutenai (aka Ktunaxa) have been proposed, but these have been unconvincing, and due to the time depth involved would probably be impossible to prove even if true. It’s also possible that the extinct Beothuk language of Newfoundland was an Algonquian language or related to Algonquian. Some researchers have pointed to potential cognates, but the quantity and quality of the Beothuk documentation is so poor and unreliable that it’s unlikely we’ll ever know the answer with certainty.

Based on the level of diversification of the languages, and in a few cases apparently on glottochronological calculations, most Algonquianists agree that Proto-Algonquian was spoken some time around 1,000-500 BC; a divergence date of ≈2,500 years ago seems reasonable enough from an impressionistic standpoint, though there are some languages which are remarkably conservative, and I think a slightly later date, say 2,000 years ago, is quite possible.

Proto-Algonquian’s exact Urheimat is uncertain. On the basis of numerous reconstructed animal and plant names, Frank Siebert in the 1960s originally placed it in the region between Lake Huron and Lake Ontario (Siebert 1967b). However, some of the species he used had extremely wide distributions and were of little use, some of his terms could not be confidently assigned to a specific species in PA, and in many cases the precise distribution of a species—especially animals—two to three thousand years ago is not known with certainty. A decade later Dean Snow revisited Siebert’s work and threw out the more problematic cases; what he was left with, however, was a region encompassing “the Great Lakes lowlands (apart from those of Lake Superior), the St. Lawrence lowlands, New England, and the Maritime Provinces” (Snow 1975:342). Another issue is the sort of circularity in these types of reconstructions, as Goddard (1994b:207) observes. For example, some Algonquian languages have a word for “harbor seal (Phoca vitulina),” reconstructable as *a·rkikwa—but of course, only those languages spoken by people who live(d) in areas with harbor seals have such words. There is certainly no word in Blackfoot or Arapaho or Shawnee for “harbor seal.” The term could have diffused in the post-PA period among people who were suddenly entering a region containing a new animal species. In any event, Siebert’s conclusions have now been abandoned, and Snow’s are of little use in determining the precise Urheimat (which was part of his point).

On the basis, instead, that Algonquian languages exhibit a series of innovations as one moves eastward, with more archaic features found in western languages, and the most divergent language, Blackfoot, found on the western Plains, Ives Goddard (1994b) argued for a homeland further west. Combined with archaeological evidence that Wiyot and Yurok speakers moved into their modern locations in California from the Columbia Plateau within the last two millennia and very strong evidence of significant contact between Algonquian and Kutenai (and to a lesser extent of non-trivial contact between Algic and Salish, Wakashan, Chimakuan, and other Pacific Northwest/Plateau languages and cultures), this suggests that Proto-Algic was also spoken in the Plateau culture region—roughly, modern Idaho, eastern Oregon and Washington, western Montana, and southern British Columbia—and Proto-Algonquian somewhat to the east of it, given the likely precontact location of Blackfoot and Kutenai speakers here. At least Proto-Core Algonquian (Algonquian minus Blackfoot) is still likely to have been spoken somewhere near the Great Lakes, though probably on the western edge.

Several attempts to identify Proto-Algonquian speakers in the archaeological record have been made. I’m not an archaeologist and know fairly little about archaeology, so it’s a bit difficult for me to analyze these attempts, but in general I’m suspicious. There seem to be two main theories. One (e.g., Denny 1989, 1991) identifies Proto-Core Algonquians with the Red Ocher and Glacial Kame burial traditions of the Midwest, and views these as direct descendants of the Western Idaho Archaic burial complex of southern Idaho, which would represent Proto-Algic or pre-Proto-Algic speakers. The other (e.g., Fiedel 1990, 1991, 2013, 2020) identifies the Proto-Algonquians with the Point Peninsula culture (or more broadly with the “Lake Forest Middle Woodland” people, including the Saugeen and Laurel cultures), and Proto-Algic speakers with some of the Shield Archaic people. Denny’s views seem to have been more widely accepted, but I see significant problems with both theories—although I don’t feel knowledgeable or competent enough to go into much detail yet.

It’s reasonable to conclude, at least, that Proto-Algic was spoken in the Plateau region, and Proto-Algonquian somewhat east of there—any location between the western edge of the Great Lakes and the Northern Plains just east of the continental divide is conceivable. Blackfoot speakers either remained in the PA homeland or later migrated back somewhat westward to their later location, while remaining Algonquian speakers migrated further east into the Great Lakes region. However, exactly when these dispersals and fragmentations took place, where each group of speakers of late dialectal Proto-Algonquian or early post-PA were located at a given time, and which archaeological cultures/complexes they may be identified with, I believe at this point remain unsettled questions.

Proto-Algonquian Phonology

Proto-Algonquian is the best-researched and most securely reconstructed protolanguage in North America, and one of the best in the world. The founding fathers of comparative Algonquian were Truman Michelson and especially the preeminent American Structuralist Leonard Bloomfield. Bloomfield initially made a reconstruction of the phonology of what he termed “Primitive Central Algonquian” in 1925, based on the four languages for which he had the most information and documentation at the time: Plains Cree, Menominee, Meskwaki (“Fox”), and southern Ojibwe (Bloomfield 1925). Famously, one of Bloomfield’s stated goals in the reconstruction was to demonstrate that languages operated and changed in the same ways in every part of the world, and consequently the classic Neogrammarian comparative method could work on unwritten, “exotic” languages just as well as on Indo-European ones, which some linguists at the time, including the great comparativist Antoine Meillet, still doubted (Bloomfield 1925:130, n. 1):

I hope, also, to help dispose of the notion that the usual processes of linguistic change are suspended on the American continent . . . . If there exists anywhere a language in which these processes do not occur . . . , then they will not explain the history of Indo-European or of any other language. A principle such as the regularity of phonetic change is not part of the specific tradition handed on to each new speaker of a given language, but is either a universal trait of human speech or nothing at all, an error.

Since he did not at the time have sufficient information on many other Algonquian languages, he did not claim his reconstruction to represent Proto-Algonquian. However, Michelson (esp. 1935) showed that by luck Bloomfield had used four languages that taken together preserved almost all the phonemic contrasts of Proto-Algonquian, and thus the reconstructions he set up could be used to derive the other daughter languages, including the superficially highly divergent Cheyenne and Arapaho (although Siebert 1941 did show that Eastern Algonquian languages retained evidence of a consonant cluster distinction not found in Bloomfield’s four languages).

Bloomfield (1946) was of lasting impact. A book chapter today referred to simply as “the Sketch,” it was a 44-page revised reconstruction of what Bloomfield now considered to essentially represent “Proto-Algonquian,” including its phonology, derivational morphology, and some inflectional morphology, and still based primarily on the same four languages, although now with recourse to an Eastern Algonquian language when necessary to distinguish some clusters. In spite of some erroneous forms and the fact that there was much that Bloomfield was still unable to account for—in particular, many aspects of the verbal inflectional system didn’t match in the languages he was comparing, and it wasn’t until the work on the Algonquian verb initiated by Ives Goddard (1967a) that these began to be resolved—the Sketch has been incredibly influential. The framework Bloomfield laid out, the terminology, most of the phonological reconstruction, and a number of aspects of the derivational and inflectional morphology have continued to the present, with modern comparative work on Algonquian, other than work on the verbs, largely consisting of modest updates or gradual improvements to Bloomfield’s work.


Proto-Algonquian can be reconstructed with the following consonants:

Proto-Algonquian Consonants

A glottal stop also existed, but only in clusters (see below). It is useful to distinguish the semivowels from true consonants (all the others).

These segments have their IPA vowels except for: č = *[tʃ], š = *[ʃ], and y = *[j]. *r was probably a lateral *[l] or a rhotic tap/flap *[ɾ], or perhaps optionally both. <l> = *[l] was its traditional reconstruction basically until Goddard (1994b) and some Algonquianists still write it as <l>.

The phonetic value of the consonant written <θ> is uncertain. Bloomfield (1946:87) rather infamously referred to it simply as an “unvoiced interdental or lateral?” with a query. The most likely possibility is the second of Bloomfield’s suggestions, *[ɬ], but also possible is something like *[ɾ̥]. Goddard and a few other Algonquianists still reconstruct it as genuinely phonetically *[θ], but I don’t think this is too widely followed. Conversely, several Algonquianists over the years have more directly claimed or argued for its status as *[ɬ], and a few wrote or write it as <ł>.

did not occur before *i(·) /i(ː)/ or *y, where it was replaced by . Except in some consonant clusters, it merged with the reflex of *r in all daughter languages except AGV and Cree-Innu, and there is some other weak evidence suggesting these two sounds were quite similar. In the languages which do keep it distinct from the reflex of *r, merged with as /θ/ in Arapaho-Besawunena and as /t/ or /ts/ in Gros Ventre, and merged with *t as /t/ in Cree-Innu, but older records show that the reflex of / was originally an affricate [tθ] in AGV, from Proto-AGV *ts.

My personal opinion is that was most likely *[ɬ] and *r was most likely *[l] or *[l ~ ɾ], given both their reflexes in daughter languages and their sources in PAc, but that the product of their merger in most daughters quickly became primarily *[ɾ]. For more detailed discussion on the value of *r and especially , including the possibility that may have had an optional affricate pronunciation *[tɬ], see here.

Since only two language groups keep them distinct, there are some PA etyma where the identity of a phoneme as *r versus is unknown; in these cases the segment is conventionally written <*L>.

Pentland (e.g., 1991:27, n. 5, 1999:226, 2006:163) reconstructs an additional consonant *kw, distinct from *k+w sequences, to explain the differing morphophonemic treatments which these sequences undergo. For example, there are two realizations of underlying *|kw-a| sequences across morpheme boundaries, *kwa and *ko·, the former of which Pentland reconstructs with *kw (e.g., *a·kema·xkʷ-aki*a·kema·xkʷaki “white ash (Fraxinus americana) trees” vs. *ka·wa·ntakw-aki*ka·wa·ntako·ki “white spruce (Picea glauca) trees”; on this see also Rule 12c in the list of (morpho)phonological rules below). There are some other minor differing treatments as well. These could be reflections of two separate phonological rules operating at different times in the pre-PA period, and it’s not always possible to distinguish a putative *kw phoneme from a *kw sequence, so I will follow the general practice and just write *kw in all instances.

is included in the table above, but it probably was not a distinct phoneme in Proto-Algonquian, although its reflex has become phonemic in the daughter languages. In PA it was in complementary distribution with *t, occurring before *i(·) and *y as well as as a variant of *t in words with diminutive consonant symbolism. *t occurred elsewhere. There are only three potential exceptions: two onomatopoeic forms—one of which also involves diminutive symbolism—and another case found in only one low-level subgroup. First, the name for “blue jay (Cyanocitta cristata)” has been reconstructed as the obviously onomatopoeic *ti·nti·wa, which has irregular reflexes in various languages, often with vowels that imply descent from PA *e· rather than *i· (e.g., Miami-Illinois teentee[kihsa] “blue jay,” teentia “plover, killdeer” [Costa 1992:27] ← Old Illinois *teenteewa), and Pentland (1983:382) suggests the alternative reconstruction *te·nt(ay)e(hs)iwa, though this has its own problems. It is also possible that the word represents a loan: compare Dakota teténiŋča, Catawba tínde, and Proto-Northern Iroquoian ≈*triʔtri:ʔ (→ Mohawk terì:teri’, Oneida tlì·tli, Onondaga díhdih or di’di·’, Cayuga dí:di:’, Seneca dí’di:’) (Siebert 1967a:53; Mithun 1984:268; Koontz 1990:12-15; Julian 2010:545-546, #450). However, what these words really suggest is just how iconic/memetic the names for this bird often are, and how such words can be distorted by remodeling them towards—or by their failure to undergo some normal sound changes in order to preserve—a certain iconic prototype that may also be shared with neighboring societies (cf. e.g. Koontz 1990:8).

There is also an onomatopoeic verb/particle *čaʔm(o·nk)- ~ *taʔm(o·nk)- “splash” (traditionally cited as *čap(o·nki), but see Pentland 1983:379-381) where the can be explained as a result of diminutive consonant symbolism. And finally, Goddard (1990:469, n. 47) argues that the Munsee TI Final -əchas “burn, heat sth.” (consisting of a Medial -əchee- “compact object; body, shape” plus the TI Final -əs “act on by heat”) must continue PA *-ečas, not something like x-ečyas, since the sequence *-ya- did not occur between consonants in PA. But this complex Final could have been subject to analogical rebuilding during its evolution from PA to Munsee, so I don’t see the need to project its shape back to the protolanguage. (Goddard 2001b:192 seems to walk this back and reconstructs an exceptional PA *-ečyas-, with interconsonantal *-ya-, in this case.)

Regardless of the above, I’ll continue to write <č> instead of <t> in Proto-Algonquian reconstructions, partly out of tradition, partly because it was phonetically *[tʃ], partly because the reflex of phonemically splits from *t in every Algonquian language, and finally because, strictly speaking, it was phonemic in PA since the very fact that it alternated with */t/ in diminutive-symbolic contexts made it contrastive with the latter, even if it did not otherwise exist as a phoneme (Nichols 1971:830, n. 5; Goddard 1979a:74).


The Proto-Algonquian vowels as traditionally reconstructed are given below. The *a-vowels are sometimes presented in charts as back and sometimes as central, but these often seem to be graphical decisions more than anything else.

Proto-Algonquian Vowels

As can be seen, the long vowels are traditionally indicated with a following interpunct/raised dot, a practice which will be continued here. Based on their realization in most daughter languages, the vowels written <e>, <e·> very likely centered around lower mid to low values *[ɛ ~ æ], *[ɛː ~ æː] (cf. Oxford 2015:320 and n. 4), <o·> certainly ranged between *[oː] and *[uː], and <i·> may well have had optional realizations as low as *[eː]; there was probably a fair amount of variation in the pronunciation of the other vowels too, given the available space and modern reflexes. There were no diphthongs, nor sequences of consecutive vowels.

Since shortly after Bloomfield published the Sketch, it has been assumed that the vowel *o was very marginal or perhaps nonexistent in PA, with /o/ or intermediate *o in daughter languages outside of Eastern Algonquian seen as deriving from PA *we sequences, except perhaps for a handful of cases of surface *o, mainly before *w (where it did not contrast with *we) or as a result of regular morphophonemic shortening processes applied to morphemes which contained *o· in other environments. This seemed bolstered by the comparative Wiyot and Yurok evidence. However, Goddard (2002a:45, n. 2, 2015b, 2018, p.c.) has recently reevaluated these conclusions, and pointed out that (1) Eastern Algonquian languages form a firmly established subgroup while the western languages do not, and (2) instead of the correspondence of non-Eastern o : Proto-Eastern Algonquian (PEA) *wə [= assumed PA *we], in his view the correspondence is (oversimplifying slightly):

  • non-Eastern o : PEA *[wə] /#__, k__
  • non-Eastern o : PEA *[oː] /elsewhere

He suggests that in fact the PA vowel was usually *o = presumably *[ʊ ~ o], which underwent a conditioned change in Eastern Algonquian (with some instances of */o/ breaking to *wə and the remaining instances merging with long *o·), while remaining unchanged in the other languages. I believe Goddard is right, and consequently I write <o> in place of traditional initial and postconsonantal <we>. More detailed discussion of this question can be found here.

A New View of Proto-Algonquian Vowels

All this said, I’ve lately come to believe that the Proto-Algonquian and Proto-Algic vowel systems were quite different from the traditional reconstructions. As this applies to Proto-Algonquian, I believe the system was organized in a manner more akin to that of PEA, and contained four full vowels and three (or four) reduced vowels: full */i/ = *[iˑ ~ e̝ˑ], */u/ = *[oˑ ~ uˑ], */æ/ = *[æˑ], and */ɑ/ = *[ɑˑ] corresponding to the traditional long *i·, *o·, *e·, *a·, and reduced */ŭ/ = *[ŭ ~ ŏ], */ə̆/ = *[ə̆ ~ ɛ̠̆], and */ă/ = *[ă] (and very marginally contrastive */ĭ/ = *[ĭ ~ ĕ̝]), corresponding to the traditional short *o, *e, *a, *i.

The Proto-Algonquian vowel system under my reinterpretation
The Proto-Algonquian vowel system under my reinterpretation. The fronted allophones of */ə̆/ are not pictured.

Under this new view, the western Core Algonquian (“Western Algonquian”) languages restructured the system by turning the full/reduced contrast into a long/short one in which each vowel was paired with a counterpart differing only in duration, including through the innovation of additional instances of *i which increased the functional load of its contrast with *i·. (At least one of these changes was also diffused to, or possibly shared with, Blackfoot.) The result was, except for the distribution of *i (and *o), the system that is traditionally reconstructed for PA, as in the chart at the beginning of the section on vowels above. PEA in some ways more directly reflected the true PA situation, though it merged the full and reduced high vowels.

This new reconstruction is based on several considerations, including comparative Algic evidence, the overall structure of the systems involved, and the most natural and economical paths of linguistic change, and is discussed in much greater detail in this post.

In spite of my new views, for the sake of continuity—and in case I’m wrong—I will continue writing the vowels as they have traditionally been written, other than the aforementioned replacement of *we by *o. And for convenience’s sake, I will usually also refer to them as “short” and “long” vowels rather than “reduced” and “full.”

Consonant Clusters

The Proto-Algonquian consonant clusters can be reconstructed as follows:

Proto-Algonquian clusters

These are clusters of true consonants, and can be referred to as true clusters. In addition, at least in theory, any true consonant or true cluster could be followed by *w or *y, though for the reasons stated previously, x(C)ty, x(C)θy, and x(C)čw did not occur, only *(C)čy, *(C)šy, and *(C)tw. (The clusters ending in *-č in the above table are parenthesized because they were allophones of the *-t clusters.) *n in the cluster *nk was phonetically *[ŋ].

The presentation in this chart differs in a few ways from tradition or the established consensus, as I will explain, while also elucidating a few other necessary points.

*r-Clusters, *hr, and *št

The row beginning with *r was reconstructed by Bloomfield using the arbitrary symbol <*ç->, as he did not know the phonemic or phonetic identity of the first member; many later Algonquianists reconstructed and wrote it as *s. However, Goddard (1994b:205) has argued, based on data from dialectal Woods Cree, that it is likely to have been *r. There are a decent number of examples of the cluster *rk, but only four examples of *rp that I know of have been proposed. Three were first identified by Pentland (1979a:65) based on the reflexes of the words in Woods Cree, and accepted by Proulx (1984b:191), who added one of his own, though Pentland himself rejected a new PA cluster as the explanation for the pattern he had found. Goddard later accepted at least two of these: *orpanih/ lung” (1994b:196, n. 14, 2015a:[14]; instead of *oxpani); and *orpenya “groundnut, Indian potato (Apios americana)” (p.c. cited by Pentland 2003:298, n. 10; instead of *oxpenya—but I think Goddard no longer considers this a valid example), though again Pentland (locis cit.) pointed to some difficulties with each of these, respectively. One issue with the supposed *rp clusters is that while they have reflexes in Woods Cree that parallel those of *rk, in other Algonquian language where a distinction is possible, the reflex is NOT parallel to that of *rk. Compare *merkwi “blood” and supposed *orpanih/ lung”: dialectal Woods Cree mithko and othpan, but Ojibwe miskwi vs. opan and Shawnee mškwi vs. hoʔpani. This alone does not firmly rule out the validity of *rp, however—cf., for instance, the reflexes of -clusters in Meskwaki: /hp/ from PA *šp but /ʃk/ from PA *šk. And Pentland (2003:298) seems open to *orpani as a valid reconstruction.

This interpretation of traditional “*çC” as *rC seems quite reasonable, but it presents some difficulties for some of the ideas advanced by Rhodes (2021). See discussion here.

Meanwhile there are two clusters given above that Bloomfield reconstructed on the basis of a single cognate set each, *hr and *št. *hr is found only in *re·hre·- “breathe (AI),” and its reconstruction is usually said to only be required by the Cree-Innu reflexes (though Blackfoot may have a unique reflex as well [Pentland 1979a:348-349; Weber 2016b:7], although I’m not fully convinced). It might thus seem more likely that the word was *re·hθe·- and developed irregularly in Cree-Innu through assimilation than that the cluster was found in just this word, a possibility suggested by Goddard (1982:26, 1994b:192). As first noted by Pentland (1979a:371), however, there is strong comparative evidence that this was originally a mimetically reduplicated form, with *re·- on its own also meaning “breathe”; the reduplication would regularly result in a cluster *-h-r-, with inserted *h. Reflexes of*re·- are now found mainly in the names for certain snakes (especially hognose snakes—cf. colloquial English names for this snake like “puff adder” and “blow(ing) snake”), but not entirely: Ojibwe newe “bullsnake (Pituophis catenifer sayi); hognose snake (Heterodon spp.),” Menominee nε·wε·w “hognose snake,” Shawnee lee- “have poison breath (AI)” and leewa “cottonmouth (Agkistrodon piscivorus),” etc. Additionally, Goddard (2013:110) later found older records of Munsee with †léewan “breath,” which would derive from an unattested verb *lee-. On this evidence, the reconstruction of a cluster *hr in this one reduplicated verb stem seems relatively secure.

Finally, a cluster *št is has been reconstructed in the single word *oštekwa·nih/ head” (stem *-štekwa·n-) although there has been much more doubt expressed over this word over the years, dating back to Michelson. It is found only in a few languages and the cluster being *št is required only by the Ojibwe and Munsee reflexes. Pentland (1977, 1979a:388) argues for its status as a loan in both, with Ojibwe oshtigwaan a clear loan from Cree oštikwān (with which I concur), and Munsee wiiləshtíikan being somehow related but in an unclear way—Pentland’s suggestion being that it represents a blend of a native Munsee prefix and an Ojibwe loan, though Goddard (1982:29 and n. 63) implicitly disputes this (viewing it as the Munsee prefix plus a direct reflex of the PA word which has been reshaped). If Pentland’s interpretation is correct, this would indicate that the actual protoform was *oʔtekwa·ni.

A remaining complication is that the Blackfoot reflex of this word has /ʔt/: mo’tokááni “head, hair” ← *me-Ctekwa·ni “a/someone’s head” (Taylor 1960:41, n. 21; Proulx 1989:63; Berman 2006:282, #102). To oversimplify somewhat, the normal Blackfoot development of PA clusters is *C1C2 → ssC2 after *i(·), otherwise → hC2 (Berman 2006:266), and so we should expect xmohtokááni (cf., e.g., -ohtoo- “place something (TI Final)” ← *aʔt-aw- “to place sth. (TI)” [Weber 2017:4]). This includes the more secure -clusters, as in *ka·šk- “sever” [initial-changed form, explained below] → ka(a)hk- (Berman 2006:271, #37). For that matter, though, it also includes numerous examples of *ʔt, and the larger point here is that this is an unexpected reflex no matter what the original cluster was: a putative *št should have been treated like other *šC clusters and *ʔt should have been treated like other cases of *ʔt, in both cases becoming ht. There is admittedly at least one case where *šk unexpectedly shows up with an initial /ʔ/ in Blackfoot, *ošk- “young, new, fresh” → o’k- “raw” (Weber 2017:6), but there are other cases which irregularly show /ʔ/ in Blackfoot for the first member of a cluster, or even where Blackfoot shows a /ʔ/+C cluster where PA just had *C. In this case, if the protoform was *-ʔtekwa·ni, then Blackfoot would at least be inheriting the cluster unchanged! At any rate, our understanding of Blackfoot historical phonology is still very imperfect and incomplete. The balance of this evidence suggests that, at best, Blackfoot cannot be a reliable source for determining if *št was a valid PA cluster.


The cluster here written *ʔm, which was not reconstructed by Bloomfield, first being identified by Goddard (1967a), is unusual in several respects and it’s not entirely clear what its actual value was. Its most common reflexes are either /m/ or whatever is the reflex of *mp in a given language, sometimes with different word-medial and word-final reflexes; but it is underlying |hm| in Delawaran. For example: PA *wi·kiwa·ʔmi “lodge” → Ojibwe wiigiwaam, Meskwaki (which has /p/ from *mp) wîkiyâpi, Plains Cree (which has /hp/ from *mp) mīkiwāhp, Munsee Delaware wíikwahm. It’s generally assumed to have been either *hm or *ʔm, and is often written <Hm> to avoid committing to one interpretation.

The primary evidence for reconstructing it with the value *hm is that in Eastern Algonquian, when the preterit suffix *-pa(n-) is suffixed to the morpheme *-eHm (PEA *-p(an-) and *-əHm) the result is *-əhəmp(an-), with an epenthetic *-e- (→ PEA *-ə-) inserted to break up the illegal cluster *Hmp, and the now intervocalic *H being realized as */h/ (Goddard 2007:249-250). Goddard suggests this treatment is also indirectly reflected in a single archaic but partially regularized form in the Miami-Illinois corpus, but the form in question doesn’t actually provide evidence on what the value of “*H” was once it became intervocalic. Because intervocalic did not occur in PA (it only occurred before consonants, while *h could occur before consonants as well as intervocalically), even if this rule can be dated back to PA times—which is not certain—then x-eʔempa could never have been a possible resolution of the sequence *-eHm-pa (Nilsen 2017:13-14). So the fact it occurred as *-ǝhǝmp in PEA unfortunately doesn’t actually tell us much about the value of *H when it was part of the cluster *Hm. Nor, for that matter, does the cluster’s value of |hm| in Delawaran, since there are no /ʔ/-clusters—or phoneme /ʔ/—in those languages. This in fact continues the PEA state of affairs, where *[ʔ] was lacking entirely; PA *ʔC clusters otherwise became *hC clusters in PEA, so even assuming the cluster in question was indeed *hm in PEA, this could derive from either PA */hm/ or */ʔm/ (or something else).

The main evidence in favor of reconstructing *ʔm is not much better, but there is some potential indirect testimony offered by Menominee; for a full discussion, see here. I actually think an old suggestion of Proulx’s (1982:396-397, n. 4) that the cluster was originally derived from the concatenation of the nominalizer *-(e)n plus an *-m, producing a geminate *mm, is fairly attractive. Even aside from the considerations which actually motivated Proulx’s proposal, *mm would be a totally anomalous/unique PA cluster—there were no other geminate clusters—which could explain its anomalous/unique developments in daughter languages, including the fact that it often shows the same reflex as *mp: *mm could very easily become intermediate *mp, in addition to intermediate *ʔm (→ *hm in some languages).


Bloomfield and later Algonquianists set up two clusters *čp and *čk. In addition to being the only occurrences of in PA outside of the rules given earlier, these clusters are very rare and basically only required by Menominee—in all other languages *čk normally has the same reflexes as *xk, though the situation with *čp is more complicated. Pentland (1979a:382-384) argues that they did not actually exist, and the modern Menominee /tsp/ and /tsk/ clusters are the result of multiple factors including vowel syncope, and diminutive consonant symbolism, assimilation/dissimilation, or other irregular processes applied to the PA clusters *xp and *xk, or sometimes *šp and *šk, or other clusters. To use the Initial "*kečk-" “elderly” as an example, most languages are consistent with *kexk-; while Menominee has kecki·w “s/he is old” and a few similar forms which motivated the reconstruction with “*čk,” there’s actually a potential Menominee doublet with the original *xk, which if true would indicate that /tsk/ is by sound symbolism: hki·weqnε·sew “s/he speaks archaically.”

However, because the ultimate provenance of a given “”-cluster is often unclear, I will continue to write traditionally reconstructed *čp and *čk as such when appropriate.

*x- and *ʔ-Clusters

The -initial clusters have been recognized as such since Bloomfield. Although */ʔ/ was not otherwise a phoneme in PA, these seem to have indeed been phonetically *[ʔ]+consonant; as exemplified below, the represents the neutralization of (pre-)PA *p, *t, or *k(w) before another true consonant. I originally treated the clusters Bloomfield wrote “*xp” and “*xk,” again using an arbitrary initial symbol, as part of the *ʔC series (following Pentland 1979a:380, 382; cf. Oxford 2016:5, 2020a:508 and Nilsen 2017:12, 2019 passim). This was based on four primary considerations. First, they are in complementary distribution with each other: in the traditional reconstruction, there are no *ʔp or *ʔk, and no *xt, *xθ, etc. Second, the reflexes of both series pattern identically in every daughter language except Menominee (which reflects the traditional -clusters with initial /ʔ/ and the traditional *x-clusters with initial /h/) and Proto-Cree (which has *št from *ʔt but *s- from *x-). Third, the initial segment is reflected as /ʔ/ in pre-Arapaho-Gros Ventre and in Cheyenne. And fourth, the neutralization process mentioned above creates both the traditional -clusters and the traditional *x-clusters. Some examples (from Pentland 1979a:373-376, 380-382; Goddard 1979a:92, 2015b:406, n. 91):

  • *ap- “sit; located” + *-tII Final” → *aʔte·wi “it is in place”
  • *nekot(-w)- “one” + suffix *-ami**nektami*neʔtami “first”
  • *pere·wa “*bird [→ later ‘fowl’]” → noun Final **-pre·w*-ʔre·w “bird” (e.g., *na·pe·ʔre·wa “male bird”)
  • *name·kw- “*fish [→ later ‘lake trout (Salvelinus namaycush)’]” + *-hs-DIM” → *name·ʔsa “fish”
  • **ši·p- “duck” → reduplicated **ši·i·p-*ši·ʔši·pa
  • *ot- “pull” + *-pw “act by mouth” → *oxpwe·wa “s/he smokes a pipe”
  • *wa·p- “bright, light” + *-kam- “water” → *wa·xkam- “bright, clear”
  • *mi·ka·t- “fight (TI)” + *-kye·ANTIP:AI” → *mi·ka·xkye·wa “s/he fights people”

The comparative Algic evidence also supports this. Wiyot and Yurok /t/ and /k/ before a consonant correspond to both Proto-Algonquian *ʔC and *xC clusters, just as (pre-)PA underlying *t-+stop and other stop+stop sequences do. For example:

  • PA *axkehkwa “kettle” : Yurok tkek’w[eˀl] ~ tkek’w[eˀr] “pot” ← PAc **tkəˀkw-
  • PA *o(·)xkw[e·wa] “maggot” : Wiyot yutw “maggots” (← pre-Wiyot *yutkʷ) : Yurok ˀye[ł] “maggot, worm” (← pre-Yurok *ytkʷ-) ← PAc **ywətkʷ-
  • PA *neʔθ[w]- “three” : Yurok nahkṣ- : Wiyot ḍikh-PAc **nəks-

There is good reason, then, to think that both PA *xp/*xk and the PA *ʔC clusters were originally a single cluster series. Contrary to my own earlier conclusions, however, there are serious problems with assuming that the original series represented *[ʔ]+consonant, or that it still represented a single series by the actual time period of Proto-Algonquian.

Most seriously, “*x” behaves like some sort of anterior fricative or other continuant in several instances. For example, Ojibwe and some other languages have irregularly changed *x- to /ʃ/ in a few words through assimilation with a neighboring-syllable /ʃ/ (e.g., *nexkanšya “my nail, claw” → Ojibwe nishkanzh, Meskwaki neshkasha, Shawnee niškaša where the regular outcomes would be xnikanzh, xnehkasha, and xniʔkaša; and *nexkenši “my forehead” → Ojibwe nishkinzh “my snout, muzzle,” where the regular outcome would be xnikinzh); this can be compared to the assimilation of /s/ to /ʃ/ in a following syllable, which was a regular sound change in the history of Ojibwe, Meskwaki, and Cree. These processes are difficult to explain under any hypothesis in which *x was a glottal stop.

Another piece of evidence, albeit indirect, for “*x-” representing *[s-] at some point is found in Southern Core Central. Meskwakian and Shawnee, like all non-Eastern languages other than AGV, merge the *θC series with the *xC series, presumably a change which happened relatively early in most of the languages. However, they both show, through diminutive consonant symbolism in the word for “girl” (which originally contained *θk) that the first consonant of the cluster must have been *[s] at some point in their history (Goddard 1994b:195, 206, 2002a:60, n. 36). Thus, Meskwaki has ishkwêsêha “girl” vs. ihkwêwa “woman.” The latter directly continues PA *eθkwe·wa, and the former includes the Meskwaki diminutive suffix -êha. The remainder can be accounted for by beginning with the PA word *eθkwe·hsa “girl” (“woman” + the PA diminutive suffix *-hs-), which has descendants in many Algonquian languages; this gave the probable intermediate form “*ixkwe·hsa” = */iskwɛːhsa/ (when *θk/*θp collapsed with *xk/*xp), and then a regular process of diminutive consonant symbolism changed the first *[s] to [ʃ]. A similar situation is found in the Shawnee cognates.[7]

A final piece of evidence for *x- being a fricative comes from Cree-Innu: the Proto-Cree reflexes of *xp and *xk are *sp and *sk. Given all this, I now follow Goddard in reconstructing *xp and *xk as Proto-Algonquian *sp and *sk, though for a few reasons I will continue to use the traditional notation for them. See here for further discussion as well as a sketch of how the original unitary cluster series in pre-PA could plausibly have given the PA state of affairs with *sp (= *xp), *sk (= *xk), *ʔt, *ʔs, etc.


In order to aid comparison with other Algonquianist sources, here I summarize the correspondences between the clusters as I will write them with how Bloomfield wrote them and how other Algonquianists write or have written them:

  • Bloomfield wrote <l> for what Goddard (1994b) and some, though not all, later Algonquianists have written <r>. Given the lack of any contrast, PA *r may well have varied between *[l] and *[ɾ], and I suspect its more common value in the protolanguage was *[l], but this newer notation is nonetheless followed here. Thus, e.g., Bloomfield’s <nl> is here written <nr>.
  • The phonetic reality of the segment Bloomfield wrote using the symbol <θ> is still debated, and some Algonquianists over the years have written it <ɬ> or <ł> instead. But most follow Bloomfield’s notation regardless of their interpretation of the likely phonetics, which I do as well, though I share the view that *[ɬ] is the most likely value.
  • Bloomfield’s consonant cluster series with <h->, N- (i.e., <m-> and <n->), <š->, and <θ-> have consistently been written the same way by all succeeding Algonquiaists, the identity of the phonemes involved is not in doubt, and the notation and interpretation is continued here unchanged.
  • While Bloomfield’s <č-> series may or may not have been distinctive in PA, it has also continued to be written as such by all succeeding Algonquianists and I follow this practice, unless there’s a word where an alternative source of Menominee /ts/+C can be very confidently determined.
  • Bloomfield’s <çk> was interpreted by most later Algonquianists as representing */sk/, and was therefore written as <sk> in numerous publications between Bloomfield’s time and the 1990s; any work before Goddard (1994b) which writes <sk> in a PA reconstruction is representing this cluster. Goddard’s newer interpretation of the cluster as */rk/ has been followed by many subsequent Algonquianists, who write <rk>, as I do as well, though with reservations. (If I were starting from scratch, I might write <çk>, reverting to Bloomfield’s more neutral notation.) If the cluster in “lung” is part of the same series, it may be written <rp>, although Bloomfield did not reconstruct a <çp> cluster.
  • Bloomfield did not reconstruct a cluster with *[h] or *[ʔ] + */m/ (or a geminate */mm/), which was first identified by Goddard (1967a). Goddard originally wrote the cluster as <P> and <m(p)>, and later (1974b:322 and following works) as <Hm>, with <H> as a cover symbol to represent either [h] or [ʔ]. Many Algonquianists continue to write <Hm>, although Goddard now writes it <hm>, while Pentland writes it <ʔm>. I follow Pentland and write <ʔm>.
  • Bloomfield reconstructed the -series as used here, but varied in how he wrote it; in 1925 he wrote this series with initial <’> while in the Sketch he used <q> instead of <ʔ> for typographical convenience. These have been written by all following Algonquianists as <ʔ> or occasionally <q> or <’>, and usually interpreted as having an initial phonetic glottal stop. I write <ʔ> like basically all other present-day Algonquianists.
  • Finally, Bloomfield reconstructed a series with initial <x>, containing *xp and *xk. Until recently, these were mostly written by Algonquianists using Bloomfield’s arbitrary symbol, except for Proulx and one or two others who wrote them with initial <t>. Goddard (1994b) reinterpreted this series as *s-initial, and it is now written by him and some other Algonquianists with <s>. I agree with Goddard’s interpretation, but I have retained Bloomfield’s more neutral notation using the arbitrary symbol <x>.

Note that I also reconstruct, and thus write, some of the PEA clusters and segments differently than traditionally. One minor difference is that PEA */kʷ/ is usually written <kw> even though it’s clear that it was a distinct phoneme; I usually write it <kw>, though occasionally I won’t for typographical reasons.

More significantly, I use <ĥp> and <ĥk> for the PEA reflexes of PA *xp/(*rp) and *xk/(*čk), versus <hp>, <ht>, <hč>, <hk> as the reflexes of PA *hp, *ht/*ʔt, *hč/*ʔč, *hk. Goddard (1994b and later) instead writes the former as <sp> and <sk> in PEA. But all Eastern languages reflect both of these cluster sets as /hp, ht, htʃ, hk/, except that Delawaran has /h/ as the reflex of PA *hk. Obviously these have to be reconstructed as two different sets of clusters in PEA, but I prefer a transcription that reflects the fact that they both probably began in [h]-like sounds over Goddard’s transcription with *s- (though he acknowledges that phonetically they probably began in something like *[ç] rather than *[s]). Once the traditional PEA *s-clusters are dispensed with, the obvious reconstruction for the PEA reflexes of the PA *š- series is *s-, their reflex in every daughter Eastern language. However, to avoid any potential confusion with Goddard’s PEA *s-clusters, I will continue to write the reflexes of PA *šp and *šk as PEA *šp and *šk, rather than *sp and *sk.

Goddard also reconstructs the PA *θC series as being continued by PEA *xC, where here <x> is meant to actually represent phonetic *[x] or *[χ]. However, the reflex in all Eastern Algonquian languages is with an anterior fricative /s-/ or /ʃ-/, with the exception of the Delawaran languages, which do have /χ-/. Especially since Delawaran represents a genetic subgroup, I consider it very unlikely that this was phonetically a dorsal fricative like *[χ] in PEA, and I write initial <θ>, continuing the PA notation. Similarly, for the cluster Goddard reconstructs as PEA *hx (continuing PA *hθ, *ʔθ, and *hr), I write <hθ>. (Rhodes 2021 has recently come to several of the same conclusions, including that PEA continued PA *hθ/*ʔθ/*hr as *[hɬ].) In other words, the correspondences between the Goddard/traditional notation for PEA and my own—where these differ—are as follows. For a more complete table see here.

PA Source Traditional PEA My PEA Notation
*xp [= *sp] + (*rp), *xk [= *sk] + (*čk) *sp, *sk *ĥp, *ĥk
*hp, *ht + *ʔt, *hč + *ʔč, *hk *hp, *ht, *hč, *hk [same]
*šp, *šk *šp, *šk *šp, *šk [initial */s-/]
*θp, *θk *xp, *xk *θp, *θk
*ʔθ/*hθ/*hr *hx *hθ

Word Shapes and Distributional Restrictions

PA words could begin with any consonant other than *h (and usually *y), with a true consonant plus semivowel, or with a vowel, but could only end in a vowel, except that some preverbs and particles ended in *h. For a long time it was thought by most Algonquianists, following Bloomfield, that PA only permitted short vowels word-finally, but it is now clear that long vowels could appear word-finally as well.[8] True clusters could only occur intervocalically, but, as noted, consonant+semivowel clusters could appear initially. There were no one-syllable nouns or verbs (though there were one-syllable noun and verb stems), and no nouns of the shape (C)(G)V̆CV̆—i.e., two-syllable nouns with two short vowels and only a single intervocalic consonant (Berman 1992). There were several positional restrictions on vowels or semivowel-vowel sequences. For example, sequences of *w or *y plus *o(·) (*wo, *wo·, *yo, *yo·) did not occur word-initially or postconsonantally, and as previously noted, there were no sequences of two or more vowels—vowels could only occur separated by consonants or semivowels.

One major phenomenon was the neutralization of the contrast between the phonemes *i and *e in the first syllable of a word. There is good internal evidence that there originally was a distinction between pre-PA *i (sometimes—I suspect always?—from even earlier *ye) and *e in initial syllables, discussed in detail in this footnote, but by the PA period the distinction had been lost. The result of the neutralization has traditionally been reconstructed as *e. However, it has various reflexes in the daughter languages. In PEA it is reflected as (← PA or at least pre-PEA *e). Ignoring some minor complications, where a distinction is possible it has the reflexes of *i- word-initially but *-e- postconsonantally in all of the non-Eastern languages including Blackfoot. As a result, Goddard now argues for the word-initial form of the neutralized vowel in PA being *i- and the postconsonantal reflex being *-e-. In his reinterpretation, this would be another innovation shared by the Eastern languages (*i*e [→ ] /#__), since all the non-Eastern languages are either ambiguous or point to word-initial *i-. However, as discussed much more fully in the post on my reinterpretation of PA vowels, this is quite unlikely, and I will continue to use the traditional notation in which the neutralized vowel is represented as *e whether following a consonant or not.

The restrictions given above result in the fact that the first syllable of a PA word could only contain the short vowels *a or *e (or *o, which was the result of the recent collapse of *we and still morphophonemic/underlying *|we|) or a long vowel. The long vowels other than *a· were also extremely rare in absolute word-initial position. There were some additional statistical tendencies which don’t qualify as exceptionless rules as well, e.g., *r and were very rare in stem-initial position, *h was a relatively uncommon consonant in general, *s was quite uncommon, etc.

One of these tendencies potentially carries particular importance: in reconstructible Core Algonquian, initial *e- was almost exclusively followed by a consonant cluster (and only certain consonant clusters); otherwise it was extremely rare, only found in before an approximant or , mainly in demonstratives and relative roots, as well as before *-n- in some demonstratives. (It also occurred before other consonants in a few probably analogical cases.) In contrast, in Blackfoot essentially all verb stems superficially appear to begin in a vowel when non-initial, most often /i-/, and this /i-/ occurs with many stems whose Core Algonquian cognates lack an initial *e-, e.g. Blackfoot -ipon- “terminate, end, be rid of” : PA *po·n- “cease.” Goddard (2018:99-101) argues that this, along with variant allomorphs of some relative roots, indicates that there was a sound change between the time of Proto-Algonquian proper (= “Proto-Algonquian-Blackfoot”) and Proto-Core Algonquian which deleted most instances of initial *e- (Goddard’s *i-); this would then be one of the few clear-cut cases where we could identify a Proto-Algonquian feature that is continued in Blackfoot but was lost or changed in the Core Algonquian languages, helping solidify the conclusion that Blackfoot is genetically the most divergent Algonquian language, a sister to all the rest (Goddard 2018). But I don’t think the evidence supports this, and it is much more likely that the loss of initial *e- occurred prior to Proto-Algonquian, while the Blackfoot stems with i- have another explanation. For further discussion, see this appendix from the post on Proto-Algonquian vowels.


Proto-Algonquian evidently did not have distinctive pitch, although an impressive number of daughter languages have developed “pitch accent” or tone systems: Arapaho, Gros Ventre, Cheyenne, some varieties of Innu, Kickapoo, Maliseet-Passamaquoddy, and Penobscot, plus to a very limited extent Nipissing and Western Algonquin.[9] While this is a sizeable number of languages, it actually represents essentially two geographic areas, the Great Plains and a region encompassing northern New England northwest through parts of Quebec and eastern Ontario. It’s probably relevant that in both of these regions there are other “pitch accent” languages (Caddoan and Iroquoian). As for the stress system, a large number of daughter languages reflect an iambic pattern in which every syllable with a long vowel and every even-numbered syllable in a sequence of two or more syllables with short vowels, counting from left to right, is prominent/stressed. The possible metrical feet in this system are (S = short, L = long, underline = prominence): (S.S), (S.L), and (L). Either final feet, or final short vowels not otherwise contained in a well-formed foot, are also stressed or are extrametrical in one way or another. This system or a very similar one, or evidence that such once existed, can be found in Abenakian, Delawaran, Menominee, Miami-Illinois, Ojibwe-Potawatomi, and at least some SNEA languages, and almost certainly represents the PA system. It is continued mostly unchanged in Ojibwe for instance, as in: ( “you (pl.) cause trouble” ← PA *keneʔšiwana·taxkamikesiʔmwa·, presumably *(ke.neʔ)(ši.wa)(na·)(tax.ka)(ʔ.mwa·) (example modified from Pentland 1992:1). While the locations of the stressed syllables are clear enough, the languages differ over which syllable takes primary stress, so the location of primary stress cannot be confidently reconstructed.[10]

Inflectional Morphology

The inflectional morphology of Algonquian languages, particularly of verbs, is extremely complex, and this section will make no attempt to come anywhere close to a complete account of it. I will discuss only those features that are most basic and important, and that will come up in later posts on the evolution from Proto-Algonquian to its daughter languages.


PA nouns distinguished two genders, animate and inanimate; two numbers, singular and plural; and proximate from obviative third persons. The “animate” gender, in addition to all humans, animals, and supernatural beings, included some notionally inanimate objects: heavenly bodies; many trees; some other plants; some items of daily use, human manufacture, and cultural importance; some body parts; etc. All other nouns were grammatically inanimate, which was the unmarked gender.

The proximate and obviative [note: the term is “obviative,” NOT “obviate”!] system was a way of distinguishing multiple animate third persons in various contexts. Any given animate noun was proximate by default, but in any clause only one animate noun could be proximate, with all others being marked as obviative; sometimes the assignment of which participants were which carried over multiple clauses. Obviation assignment was normally due to discourse-pragmatic considerations: the proximate participant represented the most topical, important, and salient one, and the one whose point of view was being featured, while obviative participants were pragmatically backgrounded. There were a few situations in which obviation assignment was mandatory, however, most notably when an animate noun was possessed by another third person, in which case it could not be proximate. Inanimate nouns were not marked for obviation, but could be covertly obviative in some contexts, with an associated verb agreeing in obviation with them.

The basic nominal suffixes of note were: AN.PROXsg *-a, AN.PROXpl *-aki, AN.OBVsg *-ari, AN.OBVpl *-ahi, INANsg *-i, INANpl *-ari. To save space, I normally gloss the animate proximate and obviative endings without also specifying that they are animate: “PROXsg” (for “AN.PROXsg”), “OBVpl” (for “AN.OBVpl”), etc.

The following table illustrates, using the nouns *eθkwe·w- “woman” (animate) and *očye·pihk- “root” (inanimate) in PA and Meskwaki:

Proto-Algonquian Meskwaki
Animate Stem *eθkwe·w- ihkwêw-
PROXsg *eθkwe·w-a ihkwêw-a
PROXpl *eθkwe·w-aki ihkwêw-aki
OBVsg *eθkwe·w-ari ihkwêw-ani
OBVpl *eθkwe·w-ahi ihkwêw-ahi
Inanimate Stem *očye·pihk- ochêpihk-
INANsg *očye·pihk-i ochêpihk-i
INANpl *očye·pihk-ari ochêpihk-ani

Nouns could be inflected for possession, with the person prefixes *ne-1,” *ke-2” (including inclusive 1pl), *o-3” (underlying *|we-|), and *me- “unspecified possessor,” as well as suffixes marking plurality and clusivity of the possessor (*-na·n1pl.EXCL,” *-naw1pl.INCL,” *-wa·w2pl/3pl”); the possessor plurality suffix preceded the gender/number/obviation suffix. As already noted, any animate noun possessed by another third person was obligatorily obviative; when the possessor was also obviative, the possessum took an additional suffix *-iriw marking this. There was a distinction between dependent nouns (inalienably possessed nouns, which always required possessive marking: primarily body parts and kinship terms), and non-dependent nouns = alienably possessed nouns. Most non-dependent possessed nouns took the “possessed theme” suffix *-em immediately following the noun stem, although there were some exceptions.

Nouns could also take the locative suffix *-enki, which in addition to its locative meaning marked the complement of a relative root (for a brief explanation of relative roots, see here); the vocative singular *-e or vocative plural *-etwike; and the diminutive *-hs, and some variants of this.

Finally, there was a special set of absentative endings which replaced the normal gender/number/obviation endings to indicate that the noun referred to an entity that was “absent, dead, departed or departing, or further removed” (Goddard 2018:97). These were formed by replacing the final *-a and *-i of the basic endings with *-a· and *-e· respectively: PROXsg *-a·, PROXpl *-ake·, OBVsg *-are·, OBVpl *-ahe·, INANsg *-e·, INANpl *-are·.


Other than the third-person pronouns, the reconstruction of Proto-Algonquian personal pronouns is reasonably straightforward. They appear to be irregular possessed forms of a root *-i·raw- (with regular loss of word-final consonants, contraction of underlying *|awe|, and similar changes):

  Singular Plural
First (EXCL) *ni·ra   |ne-i·raw| *ni·ro·na·   |ne-i·raw-ena·n|
First INCL *ki·ro·na   |ke-i·raw-enaw|
Second *ki·ra   |ke-i·raw| *ki·rawa·   |ke-i·raw-wa·w|
Third *wi·ra   |we-i·raw| (?) *wi·rawa·   |we-i·raw-wa·w| (?)

The reason I’ve querried the third-person pronouns is because the forms cited are only attested outside of Eastern Algonquian; all Eastern Algonquian languages reflect different forms, PEA *nēkəma (sg.) and *nēkəmāwa (pl.). In theory this is not very serious—we simply reconstruct PA as having *wi·ra etc., and PEA as having innovated a new set of third-person pronouns from an obscure source. The difficulty with this is that the third-person forms in the non-Eastern languages could quite easily be derived analogically from an older system which lacked them, based on the existing first- and second-person pronouns and on the pronominal inflections of dependent nouns:

“1”      *n-   :   *ni·ra
“2”     *k-    :   *ki·ra
“3”     *w-   :     X = *wi·ra

On balance, I think it’s still more likely that the Eastern Algonquian forms are innovations and those found in the non-Eastern languages reflect PA, but there is room for doubt. A fuller discussion of some of the issues and evidence involved can be found here, though that’s really not my best work and I don’t really reach any truly satisfying conclusion.

Beyond the personal pronouns, several indefinite and interrogative pronouns can also be reconstructed: *awiyaka “someone” (otherwise inflected without the *-ak-: pl. *awiyaki, OBVsg *awiyari, OBVpl *awiyahi); *ke·kw- “something”; *awe·na “who?”; *we·kw- “what?”

Finally, there were “filler” pronouns, also referred to as “noun substitutes,” PROXsg *aya, PROXpl *ayaki, OBVsg *ayari, OBVpl *ayahi, INANsg *ayi, INANpl *ayari (probably also inflectable for locative: *ayenki (?); and absentativeness: PROXsg.ABST *aya·, INANsg.ABST *aye·, etc.). To my knowledge, the usage of these pronouns in modern Algonquian languages has not been studied in much detail, with the exception of LeSourd (2003a) on Maliseet-Passamaquoddy. However, based on what is known, these were used in a way roughly equivalent to English “whatchamacallit,” to substitute for a specific noun when the speaker temporarily did not have the word they were looking for at hand (but inflected with the proper gender, number, and obviation for the context in which the noun would be appearing); when combined with descriptive Initials/preverbs, had the meaning “one with the given property, an X one” (e.g., *oški-aya or *oškaya “a young one [AN], someone young,” *oški-ayari or *oškayari “young/new ones [INAN],” etc.); as well as, likely: more broadly as a filler/hesitation marker (similar to “um…”); to introduce vagueness about a noun it was paired with (“some sort of X,” “X and so on”); and as a rough equivalent of “somebody” or “thing” when no additional specification was needed. The range of uses may have been even more complex than this, as suggested by LeSourd’s careful documentation of the various uses of the pronouns in Maliseet-Passamaquoddy, where, for instance, it is used to highlight shifts in gender, and when making corrections to earlier speech errors.


There have been two major attempts to recsontruct the PA demonstrative pronouns, which have ended up with wildly different results. These are Proulx (1988; and some later commentary by Proulx 2001, 2004b) and Goddard (2003). Pentland (2000a, 2000b) also published some limited work on reconstructing aspects of demonstratives. Proulx’s reconstruction is deeply unsatisfying and superficial, compares a number of elements which cannot be cognate, and exhibits other serious problems. Given its drawbacks and unreliability, there’s little need to go into detail on it, though I did so in an earlier version of this post.

Putting that reconstruction aside, we can turn to Goddard’s (2003) very different one. In place of Proulx’s ten demonstrative stems, Goddard reconstructs only five sets for PA, shown in their inflected forms in the table below (Goddard 2003:56; slightly modified to show long final vowels when appropriate).

Goddard’s Reconstructed PA Demonstratives
  "this" "that" "this (remote)" "that (remote)" "yonder (remote)"
PROXsg *ewa *ena *ewa·(ka·) *ena·(ka·) *eya·(ka·)
INANsg *eyo· *eni *eyo·we· *ene· *eye·
PROXpl *eyo·ki *eniki (*eyo·ke·) *ene·ke· *eye·ke·
INANpl / OBVsg *eyo·ri *eniri (*eyo·re·) *ene·re· *eye·re·
OBVpl *eyo·hi *enihi (*eyo·he·) *ene·he· *eye·he·
LOCATIVE *enahi *ena·hi *eya·hi

Unlike Proulx, Goddard presents a huge amount of supportive data as well as detailed accounts of his reconstructed history for the demonstratives in most daughter languages. In general I find his reconstructions plausible. Lacking the time to investigate all the minute details involved, it’s difficult for me to fairly evaluate it further, but on my superficial reading I do have reservations on a few points.

First, as Goddard himself more or less admits (pg. 59), the proximal “remote” (“this (remote)”) stem is not well supported for PA; it’s reconstructable for PEA but the only evidence for it outside Eastern Algonquian is a Proto-Southern Core Central adverbial particle *eyo·we “in the past.” And even if its existence in PA is accepted, the only forms that can directly be projected back from PEA are the (proximate) singular forms *ewa·(ka·) and *eyo·we·: “[t]he plural and obviative forms . . . have no direct support and have simply been assumed to match the pattern of” the distal and further distal remote stems (fourth and fifth columns).

Second and far more seriously, I don’t understand Goddard’s chosen method for reconstructing the forms in the first place, as explained on pp. 37-38. His reconstruction of the demonstratives of PEA (pp. 48-55) is very reasonable, comparing numerous languages and plausibly connecting the different forms and explaining discrepancies to arrive at a proto-version. But he doesn’t use this same method to reconstruct PA demonstratives; instead, he proceeds “inductively” (pg. 37) by comparing the PEA reconstructions only with the demonstrative systems of Meskwakian and Shawnee and deriving a “Proto-Algonquian” system from this, and then goes on to explain developments in all the other daughter languages using this reconstruction. Meskwakian is quite conservative and both it and Shawnee have a number of demonstratives, so they are certainly a reasonable place to start in trying to arrive at a PA reconstruction. But they also in the opinion of most Algonquianists, Goddard included (he calls them “closely related” and “dialectically the closest relative[s]” of each other, ibid.), form a distinct (sub-)subgroup, so we should expect that a comparison of the two will not give a full picture of the diversity of demonstratives in non-Eastern languages.[11] Assuming that this is really how he went about the task of reconstruction—rather than just being how he chose to present the data afterwards—Goddard’s approach here is methodologically unsound and raises many dangers of hasty generalization, confirmation bias, and circular reasoning, and simply (necessarily!) missing any potential demonstrative stems that happen not to have survived into Meskwakian-Shawnee and PEA.

I note with unease, for instance, that for many of the other non-Eastern languages, Goddard must posit significant phonological reductions—some well-supported and convincing, but a few stretching the bounds of credulity—and seemingly random irregular changes, to a far greater degree than occur between his reconstructed forms and those in Meskwakian and Shawnee. For example, in the evolution to Cree-Innu, the demonstratives are said to have lost word-initial PA *ey-, “and PA *e- shifts to [Proto-Cree] *a- (set B and the animate singular of set A) or is lost (set C),” while within Cree-Innu itself, “[o]ther process that affect the word-final shape of . . . demonstratives include vowel-lengthening, addition of h and a vowel, and replacement of final vowels or suffixation of -e·” (pp. 69, 73). Note that one of these changes, the fate of initial *e-, has two entirely separate outcomes depending on the demonstrative stem in question! By contrast, the main irregular phonological changes which are reconstructed to occur in the evolution to Meskwakian and Shawnee involve lengthening of the initial vowel in Meskwakian and some vowel assimilation across adjacent syllables and the loss of the initial vowel in three+ syllable forms in Shawnee.[12]

Pentland, in a couple of short papers (2000a, 2000b) discussed some aspects of the PA demonstrative pronoun system, mainly focused on the absentative/remote (“inaccessible” in his terminology) sets/endings. He accepts Goddard’s reconstruction of a system based around proximal pronouns in *ew-/*ey- and distal ones in *en-, which took either neutral or remote/inaccessible suffixes; he differs from Goddard in reconstructing some of the inaccessible endings differently, and in reconstructing a third opposition marked by demonstratives, an absentative (marked with *-iy-) alongside the proximal/distal and accessible/inaccessible oppositions. (He thus views the “inaccessible” and true absentatives as separate categories, although the PA remote pronoun stems are formally absentative.) Goddard (2003:89-93) very briefly responded to some of Pentland’s reconstructions, and I think effectively refuted the existence of absentative *-iy- and most of Pentland’s other views on this topic where the two of them differed.

Of the two major reconstructions, Goddard’s is certainly the more compelling, and his arguments generally strike me as at least plausible and often persuasive, so in future posts I will basically accept his version. Unfortunately I haven’t dug enough into the details to really insightfully judge or refute the specific evolutionary pathways he lays out. I do have one minor quibble about his section on Ojibwe (though it’s one that would actually further support his PA reconstruction), but I will reserve this discussion for the appropriate post(s) on the evolution from PA to Ojibwe.


A much more extensive discussion of Proto-Algonquian verbs, especially agreement-marking inflection, may be found in this post. Here, I will merely briefly summarize the basic categories and concepts involved, and gloss over a lot of details and caveats and exceptions.

Verbs in Proto-Algonquian were primarily inflected using one of two broad “orders,” which were syntactically conditioned “parallel sets of verb inflection . . . which express[ed] mostly the same contrasts but use[d] different morphology” (Oxford 2020a:517): the independent, generally used in main clauses, and the conjunct, generally used in subordinate clauses. There was also an imperative order for commands. The independent order was inflected with both prefixes (*ne-1,” *ke-2,” *o- = *|we-| “3,” as in nouns) and suffixes. The conjunct order was inflected solely through suffixes, as was the imperative. Several lines of evidence show that the independent is much younger than the conjunct, and ultimately derives from pre-PA possessed nouns.

Each order also contained different sub-types or modes that could provide finer-grained distinctions of aspectual, temporal, modal, or evidential meaning; these included the independent, preterit, dubitative, emphatic present, subjunctive, prohibitive, iterative, etc., as well as participles, which were formally conjunct but took nominal-like suffixes. Reconstruction of most modes has lagged behind the independent indicative.

Conjunct verbs could also undergo an ablaut process known as initial change (IC), which in PA was used at least primarily to form participles and with the iterative, and evidently in some cases also indicated perfective aspect or something similar. The pattern was as shown in the table below (Costa 1996; Goddard 2015b:368), and applied to the first vowel in the verb complex, including any preverb (a preverbal particle which expressed tense, aspect, directional, or a host of other adverbial notions).[13]

Unchanged Vowel   Changed Vowel
*a *e·
*e- *(y)e·-
*Ce *Ce·
*o *we·
*a· *aya·
*e· *aye·
*i· *a·
*o· *wa·  (or possibly *ayo·)

Verbs could also be inflected for irrealis/negation (*-w) and as diminutives (*-hsi, with the meaning “[VERB] a little bit” or indicating diminutivity of one of the verbal arguments).

PA verb stems can be sorted based on their transitivity and the animacy of their absolutive argument into four main stem classes: (V)AIs (Animate Intransitives = intransitive verbs with animate subjects), (V)IIs (Inanimate Intransitives = intransitive verbs with inanimate subjects), (V)TAs (Transitive Animates = transitive verbs with animate objects), and (V)TIs (Transitive Inanimates = transitive verbs with inanimate objects). Each of these verb classes inflected differently, though the inflection of AIs and TIs was very similar.

Unfortunately, the names and abbreviations of the classes are, in my experience, very often difficult for people new to Algonquian linguistics to recall and keep track of. One bit of help to keep in mind: the elements of the names are ordered mnemonically (Dahlstrom 2013:62-63, n. 4): first the subject (animate vs. inanimate), then the verb (transitive vs. intransitive), then the object (animate vs. inanimate). So an AI verb has an A = animate subject and is I = intransitive, while a TI verb is a T = transitive verb with an I = inanimate object.

In addition to these, there were three other classes based on one of the major stems but with an argument added or subtracted: Objectless TI verbs (verbs with TI stems that took no object), Transitivized AI verbs (verbs with AI stems that could take a “secondary object”) and Double-Object TA verbs (verbs with TA stems that could take a “secondary object” in addition to a primary object). These are respectively abbreviated TI-O or OTI, AI+O, and TA+O. A secondary object was one of the three major grammatical relations of PA, besides subjects and primary objects; there were various syntactic differences between primary and secondary objects. Because Algonquian languages are “secundative,” in ditransitive verbs (TA+Os) the primary object corresponds to the English indirect object, and the secondary object to the English direct object. For example, the primary object of *mi·r- “give to (TA+O)” was the person to whom something was given, and the secondary object was the item given.

The following table summarizes the PA major and minor stem classes:

Summary of PA stem classes
“OBJ(1)” = primary object, “OBJ(2)” = secondary object. (Click to expand.)

And the following charts lay out the basic inflectional template for each verb class—ignoring elements, like the irrealis, that are not involved in agreement marking. The various slots, and how they instantiate agreement patterns, will then be summarized.
Simplified diagram of independent and conjunct order TA verb templates

Simplified diagram of independent and conjunct order TI verb templates

Simplified diagram of independent and conjunct order AI verb templates

Simplified diagram of independent and conjunct order II verb templates

In the independent order, the person prefix and central ending taken together marked one of the verb’s arguments (the “central participant”), and any peripheral ending, if present, the other (the “peripheral participant”). The peripheral participant was always third person, and always less discourse-salient than the central participant—because the central participant was non-third person, or because the central participant was proximate (or a more salient obviative) and the peripheral participant was (a less salient) obviative, or because the central participant was animate and the peripheral participant was inanimate.

For the most part, the theme sign, found in transitive verbs, indicated the person of the verb’s object. In TA verbs the possibilities were: *-i1OBJ,” *-eθ2OBJ,” *-a·3OBJ” (always animate third persons, of course, since these were TA verbs), or *-ekw “inverse.” For TI theme signs, see the following footnote.[14] When one of the normal object-marking theme signs was used, the central participant was normally the subject of the verb and peripheral marking, if present, indexed the object. If the inverse theme sign *-ekw was used, however, central agreement switched to marking the notional object of the verb, and any peripheral participant was the notional subject. *-ekw was the theme sign whenever the verb’s notional object was more salient than its notional subject, or, in the independent order only, when the subject was third person and the object was first or second person.

In the conjunct, the central and peripheral endings were a bit different: the central ending could mark any of the subject, object, or both simultaneously via a portmanteau suffix (e.g., *-ament3AN»1pl.EXCL”), and there was no peripheral ending except in participles. When the inverse theme sign was used, though, the central ending still always indexed only the object, as in the independent.

Verbs in which the theme sign marked the object as third person and the central participant was the subject can be described as showing direct agreement; those in which the theme sign marked the object and the central participant was variably the subject or object or both can be described as showing neutral agreement; and those in which the inverse theme sign was used and the central participant was the notional object can be described as showing inverse agreement. Arguably, inverse agreement in some cases indicated a voice distinction (“inverse voice” or “patient voice,” in which the notional subject and object roles were reversed) and in other cases—those with a third-person subject and first- or second-person object—was just a shallow matter of obligatory agreement morphology, with no deeper syntactic implications. See the post on Proto-Algonquian verbs for more discussion.

As can be seen in the verb templates above, the central endings were made up of several parts. In the conjunct, they consisted mainly of just a central suffix which marked the features of one or more verbal arguments; this suffix could have its features enriched with an “augment,” such as *-riMARK.OBV.SBJ” if a third-person subject was unexpectedly/contrastively obviative.[15] In the independent order, if the central participant was plural, this was indicated with a pluralizer (e.g., *-naw1pl.INCL,” *-wa·w2pl/3pl”). Finally, the independent central ending always included what I will, for convenience, analyze as a usually meaningless, submorphemic “formative” element (Goddard’s term), even though this analysis is more appropriate for a pre-PA time period (Goddard 2007:264). The four such formatives were: an “M-formative” *-eʔm, an “N-formative” *-n(ay), a “W-formative” *-w, and a formative that marked certain third persons, *-ẅ “3.” The w-umlaut notation is to indicate that unlike the homophonous W-formative, *-ẅ caused “umlaut” of a preceding *a· to *e·. Which formative was used in any given case was conditioned by the properties of the peripheral suffix (or its absence). Other than *-ẅ, I normally gloss the formatives just as “FMV,” or if it is necessary to distinguish them, as “M.FMV,” “W.FMV,” and “N.FMV.” The reconstruction of the N-formative here is potentially controversial; see the PA verbs post for discussion.

One final important contrast to note is between objective and absolute verbs, a sort of differential argument marking system which existed in the independent. If the object (or inverse subject) of a transitive verb was a definite NP or was pronominal, the objective inflection was used, which primarily involved indexing the object via the peripheral ending. If it was an indefinite NP, it was not indexed on the verb, and there was either no peripheral ending, or the peripheral ending, exceptionally, marked the same third-person subject (or inverse object) as the central ending (using the formative *-ẅ) did.

Again, for much more discussion and exemplification, including various sample paradigms, see the post dedicated to verbs. Here, to illustrate a few of the basic ideas involved, I’ll just present three sample verbs, one in the independent order, one conjunct—with the inflection selected in order to illustrate initial change—and one imperative. The first is an objective TI (= transitive with inanimate object) and the other two are TA (= transitive with animate object), all with a second-person plural subject; the independent example also includes a preverb. The various verbal elements are color-coded to help distinguish them and illustrate their role in agreement marking. Recall that the central ending (consisting of a central suffix and/or pluralizer plus, in the independent, a formative and any prefix) indexes the “central participant,” or in some conjunct and imperative verbs, both the subject and object; the peripheral suffix, if present, indexes the “peripheral participant”; the theme sign indicates the person of the object or else that the verb is “inverse” (and hence that the central participant is the notional object rather than the subject or anything else). The formative is selected based on the kind of peripheral suffix present and the properties of the argument the latter marks, and other than *-ẅ is meaningless in and of itself.

Sample independent order verb
Sample independent order verb (TI objective “hear something” with second-person plural subject and definite inanimate plural object, and a preverb).
Sample conjunct order verb
Sample conjunct order verb (TA “hear someone” with second-person plural subject and first singular object); the iterative mode triggers initial change (ablaut of the first vowel).
Sample imperative order verb
Sample imperative order verb (TA “hear someone” with second-person plural subject and first singular object).

(Morpho)phonological Rules

The following are among the most important phonological and morphophonemic rules which can be reconstructed back to the Proto-Algonquian period. The list is not exhaustive, and some of the rules as stated leave out some details if they are relatively minor or otherwise aren’t pertinent for our purposes. In some cases there are also alternative ways of formulating the suite of rules to produce the same output. Although in some cases the order doesn’t matter to give the correct output, many rules are consciously ordered with respect to one another. Several rules, in particular those involving contraction of vowel-glide-vowel sequences (VGV, i.e. *VwV or *VyV), and especially those involving the sequence *|awe|, are still rather unclear and have yet to be fully worked out for PA. These VGV contractions specifically will be addressed in a separate section below.

In the list of changes, I’ll first provide a brief prose summary of the change, then a formal notation, followed by an example or examples in both PA and modern languages when possible.[16] In some instances I will leave out some grammatical suffixes that are irrelevant to the rule at issue from the listing of morphemes in the input of a given example, but this should hopefully not cause any confusion.

In the formal notation, “#” indicates a word boundary, “+” a morpheme boundary, “C” a consonant, “N” a nasal consonant, “V” a vowel, “Vː” a long vowel, “G” a semivowel, “Ø” null, “≠” except in the defined environment, and “α” a given value out of a set of possible values; “POA” stands for “place of articulation.” There is also a special morphophoneme |ẅ|. Finally, note that surface *o in Proto-Algonquian was morphophonemic/underlying *|we| in most cases, so when appropriate I treat it as such here and write *we in the input to these rules even when the the output contains *o. More properly, I should treat this *|we| → *o realization as just another of the rules, but I don’t currently have time to rewrite this whole section to accommodate this.

(And yes, I realize these sorts of rules have not been in vogue in linguistics for over two decades, but whatever, I’m not writing up an OT treatment of Proto-Algonquian. Not a big fan of most of OT anyway.)

(1a) An epenthetic vowel, either *-e- or *-i- (known as “connective *e” and “connective *i”), is inserted to break up most consonant clusters across a morpheme boundary. According to Pentland (1999:246) the vowel was *-i- before Medials and *-e- elsewhere, while Proulx (1977, 1984b:197-198, n. 11, 1985:61) argues that it was *-i- in primary derivation and *-e- in secondary derivation and inflection, though he admits (1977:156, 1984b:198, n. 11) that this account “leaves some irregularit[ies]” in “a few archaic combinations”; Goddard (1990b:451, n. 9) simply denies that the environment where one epenthetic vowel was chosen over the other is fully predictable. *-e- is the only one found in inflection, though, with one possible exception. It’s often difficult to determine when Rule 1a actually operated, because in many cases it’s hard or impossible to choose whether to segment a given morpheme as, say, *e-initial, or as *C-initial, which inserts epenthetic *-e- following consonants by Rule 1a. There are sometimes clues which help solve the question, though. Rule 1a must precede Rules 4, 9, 12a, 16b, and 18. However, in certain individual cases it must follow Rule 13.
Ø → e, i /C+__C (morphologically/lexically conditioned)

  • *šekwah- “crush by tool (TI)” + *-ke·ANTIP:AI” → *šekwahike·- “crush (things) by tool (AI)” (e.g., Ojibwe zhishigwa’ige [reduplicated], Munsee škwahiikee- [O’Meara 1990:194]); compare, showing that the *-i- of *-i-ke· is epenthetic: *ahšam- “feed, give food to (TA)” + *-ke·*ahšanØke·- “feed (people), serve a meal (AI)” with no epenthetic vowel (e.g., Ojibwe ashange, Munsee xangee- [O’Meara 1990:203]).
  • *nep- “die (AI)” + *-t3AN.CONJ” + *-e·SBJV” → **nepte·**nepke· [*-t*-k after a consonant, by Rule 13 below] → *nepeke· “if s/he dies” (e.g., Meskwaki nepeke, Menominee nepε·k); epenthetic status of *-e- is confirmed by the postconsonantal allomorph of the following consonant, and cf. as well, with no epenthetic vowel: *nep- + *-ẅ3” + *-aPROXsg” → *nepØwa “s/he dies” (e.g., Meskwaki nepwa, Menominee nepuah).

(1b) Rule 1a failed to apply with certain specific morphemes and morpheme combinations. To take one example, epenthetic vowels were not added between the consonant of an inflectional morpheme and the 3AN.CONJ suffix *-t ~ *-k; for instance the 2OBJ theme sign *-eθ was added directly to *-t ~ *-k*-eθk, as in *wa·pameθki “when s/he looks at you.” Rule 1a also failed to apply in some cases of derivation and reduplication in (pre-)PA, leading to many PA *ʔC and *xC clusters, as noted previously.

(2a) The *e of the person prefixes (*ne-, *ke-, *we-, *me-) is lost before a vowel-initial dependent noun stem. Rule 2a must precede Rules 2b, 5, and 11.
e → Ø /#{n,k,w,m}__+V (morphologically conditioned)

  • *ne-1” + *-i·yaw- “body” → *nyawi “my body; myself” (e.g., Miami-Illinois niiyawi)
  • *ke-2” + *-atay- “belly, stomach” → *katayi “your belly, stomach” (e.g., Plains Cree katay, Munsee kătay)

(2b) Before all other vowel-initial stems (= verbs and non-dependent nouns), an epenthetic *-t- is inserted after the person prefixes. Rule 2b must follow Rule 2a and must precede Rule 5.
Ø → t /#{ne,ke,we}+__V (morphologically conditioned)

  • *ne-1” + *-awasw “warm self (AI)” → *netawaso “I warm myself” (e.g., Ojibwe nindawaz)
  • *ke-2” + *-ehkw “louse” + *-emPOSSESSED.NOUN” → *ketehkoma “your louse” (e.g., Plains Cree kitihkom)

(3) The 3OBJ theme sign *-a· is deleted before a vowel-initial agreement suffix in the conjunct or imperative. Rule 3 must precede Rule 5.
a· → Ø /__+V (morphologically conditioned)

  • *wa·pam- “look at (TA)” + *-a·3OBJ” + *-ak1sg»3AN.CONJ” → *wa·pamØake· “if I look at h/” (e.g., Ojibwe waabamag “if I see h/”); compare, with intervening suffixes protecting the theme sign from deletion, *wa·pam- + *-a· + *-hsiDIM” + *-wIRR” + *-ak*wa·pamhsiwake· “if I don’t look at h/ a little bit” (e.g., Ojibwe waabamaasiwag “if I don’t see h/”)
  • *wa·pam- + *-a· + *-ehko2pl»3AN.IMPER” → *wa·pamØehko “(you pl.) look at h//them!” (e.g., Meskwaki wâpamehko, Plains Cree wāpamihk “(you pl.) see h//them!”)

(4) The formative *-n(ay) has the allomorphs *-(e)ne·- before a consonant (= before one of the pluralizers or before a consonant-initial mode suffix: by Rule 9), and *-(e)n- before a vowel. (It surfaces with an initial *e when following a consonant—by Rule 1a—and with no initial *e when following a vowel, including the *a· allomorph of the TI(1) theme sign *|am| [see Rule 16a].)[17] Rule 4 must follow Rules 1a and 1b and must precede Rule 5, and as noted is partly encompassed by Rule 9.
n(ay) → n /__+V
n(ay) → ne· /__+C (morphologically conditioned)

  • *we-3” + *meθk- “find (TI)” + *-am [→ *-a· by Rule 16a] “INAN.OBJ(CL1)” + *-n(ay)N.FMV” + *-ariINANpl” → *omeθka·nari “s/he finds them (INAN)” (e.g., Munsee ‡moxkaḿənal, Ojibwe omikaanan)
  • *we- + *meθk- + *-am [→ *-a·] + *-n(ay) + *-wa·w2pl/3pl” + *-ari*omeθka·ne·wa·wari “they find them (INAN)” (e.g., Munsee moxkamənéewa “they find it/them (INAN),” Ojibwe omikaanaawaan with /aː/ the regular reflex of “recontracted” *|ay-e|, see below)

(5) An epenthetic *-y- is inserted between two vowels in most cases.[18] Rule 5 must follow Rules 2a, 2b, and 3 and must precede Rule 9.
Ø → y /V__V

  • *wi·ki- “dwell (AI)” + *-a·n1sg.CONJ” → *wi·kiya·ni “when I dwell” (e.g., Maliseet-Passamaquoddy wikiyan); compare *wi·ki- + *-wIRR” + *-a·n*wi·kiwa·ni “when I don’t dwell” (e.g., Maliseet-Passamaquoddy wikiwan)
  • *wa·pam- “look at (TA)” + *-i1OBJ” + *-an2sg.CONJ” → *wa·pamiyane· “if you look at me” (e.g., Ojibwe waabamiyan “if you see me,” Meskwaki ‡wâpamiyane)
  • *ašye·- “return, back” + *-o·te· “crawl (AI)” → *ašye·yo·te·wa “s/he crawls backwards” (e.g., Ojibwe azheyoode)

(6) The irrealis suffix *-w (and a derivational suffix *-ẅ, but not the formative *-ẅ [Goddard 2007:232]) inserts a preceding epenthetic *-o- when following a consonant. Rule 6 must precede Rules 8, 9, 11, 12c, 14a, and 14b.
Ø → o /C__+w (morphologically conditioned)

  • *ne-1” + *wem- “come from somewhere (AI+R)” + *-wIRR” → *no·mowe “I don’t come from there” (e.g., Unami nú·mwi)

(7a) The formative *-ẅ causes a preceding *a· to umlaut to *e·. Everyone agrees on this, but according to Pentland (1999:236, 238, 246), two other suffixes cause this umlaut as well, *-riMARK.OBV.SBJ” and the verbal (and originally nominal) diminutive *-hsi. Pentland’s evidence for the obviative subject suffix inducing umlaut strikes me as very weak, so I don’t subscribe to it here. On the other hand, umlaut before the diminutive seems possible, and I’ve given a possible example of it below. Rule 7a must precede Rules 8 and 9.
a· → e· /__+{ẅ,(hsi)} (morphologically conditioned)

  • *pya·- “come (AI)” + *-ẅ3” + *-aPROXsg*pywa “s/he comes” (e.g., Meskwaki pyêwa, Munsee peew, Menominee pw, Miami-Illinois piiwa); compare, with no *ẅ-formative to condition umlaut, *ne-1” + *pya·-*nepy “I come” (e.g., Meskwaki ‡nepya, Munsee mba, Menominee nepiam, Miami-Illinois nimpya)
  • *na·θ- “fetch (TA)” + *-a·3OBJ*-ẅ + *-a*na·θwa “s/he fetches (someone)” (e.g., Menominee na·nε·w “s/he (PROX) fetches him/her/them (OBV)”); compare *we-3” + *na·θ- + *-a· + *-wW.FMV” + *-ariOBVsg” → *ona·θwari “s/he (PROX) fetches h/ (OBV)” (e.g., Ojibwe onaanaan “s/he (PROX) fetches him/her/them (OBV)”)
  • *ne-1” + *nepa·- “sleep (AI)” + *-hsiDIM” → *nenephsi “I sleep a little bit; little me sleeps” (e.g., Plains Cree ‡ninipēsin); compare *ne- + *nepa·-*nenep “I sleep” (e.g., Plains Cree ‡ninipān)

(7b) Conversely, several morphemes instead cause umlaut of a preceding *e· to *a or *a·. These included the nominalizing suffix *-n; the formative *-n(ay); and the archaic nominalizing suffix *-(e)ʔm.
e· → a(·) /__+{n(ay),n,(e)ʔm,…} (morphologically conditioned)

  • *ki·škešike·- “cut off/through (AI)” + *-nNMLZ” → *ki·škešikani “tool for severing something” (e.g., Ojibwe giishkizhigan “scythe,” Plains Cree kīskisikan); compare *ke-2” + *ki·škešike·-*keki·škešik “you’re cutting (something) off/through” (e.g., Ojibwe gigiishkizhige, Plains Cree ‡kikīskisikēn)
  • *či·me·- “paddle (AI)” + *-nNMLZ” + *-ariINANpl” → *či·ma(·)nari “canoes” (e.g., Ojibwe jiimaanan, Meskwaki ‡chîmânani, Arapaho θííwonoʔ); compare *ne-1” + *či·me·-*neči·m “I paddle” (e.g., Ojibwe ninjiime)
  • *ne-1” + *mi·riwe·- “give (AI/TI(3))” + *-n(ay)N.FMV” + *-iINANsg” → *nemi·riwni “I give it (away)” (e.g., Penobscot nə̀miləwαn); compare *ne- + *mi·riwe·-*nemi·riw “I give” (e.g., Penobscot nə̀miləwe)
  • *wenike·- “portage (AI)” + *-(e)ʔmNMLZ” → *onikaʔmi “portage (n.)” (e.g., Ojibwe onigam); compare *ke-2” + *wenike·-*ko·nik “you portage” (e.g., Ojibwe g[id]oonige)

(8) At least in most cases, the second of two consecutive semivowels drops (though contrast Pentland 1999:248-249). Rule 8 must follow Rules 1a/1b, 6, and 7a, must precede Rules 9, 12a, 12b, and 12c, and fails to apply in the context defined by Rule 16b.
G → Ø /G__

  • *ne-1” + *-hčiw- “fleshy part of upper arm or lower leg” + *-ye·wakw “flesh” → **nehčiwye·wakwi*nehčiwe·wakwi “my upper arm, biceps”(?) (e.g., Menominee nεhce·wε·wak) [probably not of real PA date, and gender not certain; glosses and combination are from Goddard 1982:39-40, 2001b:190, n. 50]
  • *ke-2” + *koʔθ- “fear (TA)” + *-a·3OBJ” + *-wFMV” + *-wa·w2pl/3pl” + *-akiPROXpl” → **kekoʔθa·wwa·waki*kekoʔθa·wa·waki “you (pl.) are afraid of them” (e.g., Moose Cree ‡kikoštāwāwak)

(9) Many vowel-semivowel-vowel sequences are contracted. While many can be summarized as V1Ge → V1ː, a number of the details are complex, especially for certain combinations. These will be described in more detail further below. Rule 9 must follow Rules 1a, 5, 6, 7a (since its output is not subject to umlaut), and 8, must precede Rules 11 and 16b, and is partly responsible for Rule 4.
VGV → Vː (morphologically/lexically conditioned)

(10) [obsolete, but kept in list to preserve numbering]

(11) Word-initial and postconsonantal *w is deleted before *o(·). Rule 11 must follow Rules 2a, 6, and 9.
w → Ø /{#,C}__o(·)

  • *we-3” + *-o·hθ- “father” + *-ariOBVsg” → **wo·hθari [by Rule 2a] → *Øo·hθarih/ father” (e.g., Menominee o·hnan)
  • *meʔtekw- “tree” + *-o(·)θ- “boat” → **meʔtekwo(·)θ-*meʔteko(·)θ- “wooden boat, dugout canoe” (e.g., Plains Cree mistikōsi “wooden boat,” Menominee meqteks [Bloomfield 1925:139-140])

(12a) The sequence *-ye- between two consonants contracts to *-i-. Rule 12a must follow Rules 1a and 8.
ye → i /C__C

  • *wi·kopy- “basswood (Tilia americana) bark” + *-ehke· “obtain, make, use (AI)” → *wi·kopihke·wa “s/he gathers/prepares basswood bark” (e.g., Meskwaki wîkopihkêwa)
  • *nepy- “water” + *-enkiLOC” → *nepinki “in the water” (e.g., Kickapoo nepiki, Shawnee nepiki, Arapaho nečíʔ)

(12b) The sequence *-ya- between two consonants becomes *-ye·- (possibly with a couple of lexically/morphologically conditioned exceptions). Rule 12b must follow Rule 8.
a → e· /Cy__C

  • *aʔseny- “rock, stone” + *-ariINANpl” and/or *-akiAN.PROXpl[19]*aʔsenye·ri ~ *aʔsenye·ki “rocks, stones” (e.g., Ojibwe asiniin ~ asiniig, Meskwaki asenni, Shawnee [pak]aʔθenyeeki “pebbles”)

(12c) The sequence *-w-a- between consonants in nominal and verbal inflection and in derivation was usually contracted to *o·, but there were exceptions. (There don’t appear to have been any instances of morpheme-internal *waC except when preceded by *k [Nilsen 2017:20, citing p.c. from Goddard].) The pattern in nominal inflection was apparently that contraction occurred: (1) always in inanimate nouns; (2) always in animate stems ending in a consonant other than the sequence *-kw; (3) almost never (perhaps never) in names for animals ending in *-kw; and (4) sometimes, but not always, in other animate stems ending in *-kw. In verbal inflection, contraction always occurred after the inverse theme sign *-ekw, or between the formatives *-w or *-ẅ and a following peripheral suffix. Rule 12c must follow Rules 6 and 8.
w+a → o· /C__C (morphologically/lexically conditioned)

  • *meʔtekw- “stick; tree” + *-ariINANpl” → *meʔtekri “sticks; trees” (e.g., Ojibwe mitigoon “sticks,” Meskwaki mehtekôni, Arapaho bêeteíí “bows”) [inanimate noun: has contraction]
  • *we-3” + -i·θemw- “cross-sibling-in-law” + *-ahiAN.OBVpl” → *wi·θemhih/ cross-siblings-in-law” (e.g., Ojibwe dialectal wiinimoo, Munsee dialectal wíiləmool [Goddard 2010:19]) [animate noun stem not ending in *-kw: has contraction]
  • *atehkw- “caribou (Rangifer tarandus)” + *-akiAN.PROXpl” → *atehkwaki “caribou (pl.)” (e.g., Ojibwe adikwag, Munsee ătóhwak “deer (pl.)”) [word for animal ending in *-kw: lacks contraction]
  • *aθa(·)nkw- “star [AN]” + *-akiAN.PROXpl” → *aθa(·)nkwaki “stars” (e.g., Meskwaki anâkwaki, Shawnee halaakwaki, Unami alánkɔk [LTD, ← pre-Unami *alá·nkwak]) [non-animal animate stem ending in *-kw: lacks contraction]
  • *axkehkw- “kettle, pot [AN]” + *-akiAN.PROXpl” → *axkehkki “kettles, pots” (e.g., Ojibwe akikoog, Meskwaki ahkohkôki “kettles, drums,” Shawnee haʔkoʔkooki, Menominee ahkε·hkok) [non-animal animate stem ending in *-kw: this time with contraction]
  • *we-3” + *na·θ- “fetch (TA)” + *-ekwINV” + *-wFMV” + *-ariOBVsg” → **wena·θekwari [*-ww-*-w- by Rule 8] → *ona·θekri “s/he (OBV) fetches h/ (PROX)” (e.g., Ojibwe onaanigoon “s/he (OBV) fetches him/her/them (PROX)”)

(13) The animate third-person conjunct suffix has the allomorphs *-t after vowels and *-k after consonants. Rule 13 must follow Rule 1b and must precede Rules 15 and 16a. However, in certain individual cases it must also precede Rule 1a.
t → k /C+__ (morphologically conditioned)

  • *pi·ntwike·- “enter a lodge/building (AI)” + *-t3AN.CONJ” + *-e·SBJV” → *pi·ntwike·t “if s/he enters” (e.g., Ojibwe biindiged)
  • *takwihšin- “arrive (AI)” + *-t + *-e·*takwihšink “if s/he arrives” (e.g., Ojibwe dagoshing)
  • *pya·- “come (AI)” + *-t + *-e·*pya·t “if s/he comes” (e.g., Munsee páate)
  • *pya·- “come (AI)” + *-wIRR” + *-t + *-e·**pya·wk*pya·kwe· [see Rule 19 below] “if s/he doesn’t come” (e.g., Munsee páakwe, Meskwaki [mêhi-]pyâkwe “before s/he came/comes”)

(14a) *h is lost between two consonants (when Rule 1a failed to apply). Rule 14a must follow Rules 1b and 6 and must precede Rules 15 and 16a.
h → Ø /C__C

  • *wa·pant- “look at (TI)” + *-am INAN.OBJ(CL1) + *-hkPROHIB” + *-an2sg.CONJ” + *-iCONJ.INDIC” → **wa·pantamØkani*wa·pantankani [by Rule 15 = 16a] “don’t (sg.) look at it!” (e.g., Ojibwe [gego] waabandangen “don’t (sg.) see it!”)
  • *wa·pant- + *-am + *-hsiDIM” + *-w(-e)IRR” + *-ẅ3” + *-aPROXsg” → **wa·pantamØsi·wa [*-iwe-*-i·- by Rule 9] → *wa·pantansi·wa [by Rule 15] “s/he doesn’t look at it a little bit” (e.g., Miami-Illinois waapantansiiwa “s/he doesn’t look at it,” Ojibwe [gaawiin] owaabandanziin “s/he doesn’t see it”)
  • *wera·kan- “bowl, dish” + *-hsDIM” + *-iINANsg” → *ora·ke·nØsi [with archaic ablaut of the vowel before the diminutive suffix?] “small bowl, small dish” (e.g., Unami lɔ́·k·e·ns).
  • For another probable example of Rule 14a, operating in pre-PA, see *name·kw- + *-hs- → (**name·kws- →) *name·ʔsa “fish,” cited above in the discussion on *xC and *ʔC clusters.

(14b) There is at least one clear exception to Rule 14a which I’m aware of, namely that *θhs resolved to *hs, not xθs, which was an illegal cluster. Perhaps the rule more broadly was that the *-h- was deleted when a legal cluster would result, and otherwise the first consonant in a *ChC sequence was deleted, leaving a final legal *hC cluster?
θ → Ø /__hs   (/__hC ?)

  • *wa·pam- “look at (TA)” + *-eθ2OBJ” + *-hsDIM” + *-ow [← *-w by Rule 6] “IRR” + *-a·n1sg.CONJ” → **wa·pameØhsowa·ni “when I don’t look at you a little bit” [opaque due to the seeming absence of the theme sign *-eθ, and repaired by different daughters by re-inserting *-eθ:] → pseudo-PA *wa·pamehseθowa·ni (Ojibwe waabamisinowaan “if I don’t see you”) / → pseudo-PA *wa·pamehso·θa·ni [*-owe-*-o·- by Rule 9] (Miami-Illinois waapamehsoolaani “I don’t look at you,” Cheyenne [tséssáa]vóomáhetse = |-vóom-aʹhéte| “when I didn’t see you”)
  • *ne-1” + pre-PA *-šiθ- “father-in-law” [cf. *ošiθem- “have as a father-in-law (TA),” *nešiθehsa “my father-in-law”] + *-hsDIM” → *nešiØhsa “my maternal cross-uncle” (e.g., Menominee nese·hsak (pl.), Shawnee nišiʔθa [← older Shawnee *nišihsa], Ojibwe nizise [with sibilant assimilation] ~ nizhishenh [with diminutive consonant symbolism])

(15) A nasal assimilates in place of articulation to any following consonant (when Rule 1a failed to apply). Rule 15 must follow Rules 1b, 13, and 14a, and must precede Rule 16a.
N → [αPOA] /__C[αPOA]

  • *akim- “count (TA)” + *-tTI” + *-amINAN.OBJ(CL1)” + *-a·n1sg.CONJ” + *-iINANsg.PART” + IC*e·kintama·ni “what (INANsg) I count” (e.g., Ojibwe egindamaan, Miami-Illinois eekintamaani “I count it”)
  • *wa·pam- “look at (TA)” + *-t + *-am + *-ro2sg.IMPER” → *wa·pantanro “look (sg.) at it!” (e.g., Miami-Illinois waapantanto)
  • *šek- “urinate” + *-enkwa·m “sleep (AI)” + *-t3AN.CONJ” + *-e·SBJV” → **šekenkwa·mke· [*-t-*-k- by Rule 13] → *šekenkwa·nke· “if s/he wets the bed” (e.g., Ojibwe zhigingwaang)

(16a) The TI(1) theme sign *|-am| has the allomorphs: (1) *-e· before the formative *-(e)ʔm; (2) *-a· before the formative *-n(ay); (3) *-an before the 3AN.CONJ suffix *-k [← *-t by Rule 13] and 2sg.IMPER suffix *-ro; and (4) *-am elsewhere.[20] (The treatment of *|-am| when the MARK.OBV.SBJ suffix *-ri followed is not totally clear, but most likely an epenthetic vowel was inserted by Rule 1a, as discussed here.) In many daughter languages one or another of the allomorphs has been extended to additional environments. Rule 16a must follow Rules 1b, 13, 14a, and 15 and perhaps 7b and must precede Rule 17.
am → am, a·, e·, an (morphologically conditioned)

  • *ne-1” + *no·nt- “hear (TI)” + *-amINAN.OBJ(CL1)” + *-(e)ʔmM.FMV” + *-na·n1pl.EXCL” → *neno·ntʔmena· “we (EXCL) hear (something)” (e.g., Menominee ‡neno·htε·menaw “we (EXCL) hear it”; cf. Ojibwe ninoondaamin with the *-a·- allomorph analogically extended)
  • *ne- + *no·nt- + *-am + *-n(ay)N.FMV” + *-iINANsg” → *neno·ntni “I hear it” (e.g., Ojibwe ninoondaan, Menominee ‡neno·htn)
  • *no·nt- + *-am + *-k3AN.CONJ” + *-iri “iterative” + IC*nwa·ntankiri “whenever s/he hears it/them (INAN)” (e.g., Ojibwe nwaandangin, Menominee nayo·htahken)
  • *no·nt- + *-am + *-a·n1sg.CONJ” → *no·ntama·ne· “if I hear it/them (INAN)” (e.g., Ojibwe noondamaan, Menominee no·htaman “if I/you (sg.) hear it/them (INAN)”)

(16b) The precise allomorphy of TI(2) theme sign *|-aw| is more difficult to reconstruct, since none of the daughter languages apparently reflects the original distribution exactly. In general it seems to have been realized as *-o· before the formatives *-n(ay) and *-eʔm, as *-a· before the formative *-ẅ and the MARK.OBV.SBJ suffix *-ri, and as *-aw mostly elsewhere, though not all the details are known (Goddard 1979b:72-73; Pentland 1999:253-254). Before the animate third-person conjunct suffix *-t ~ *-k, the underlying allomorph was *-aw, but the *w metathesized with the following consonant and the vowel was compensatorily lengthened, giving surface *-a·kw. See here for much more on the distribution of the TI(2) theme sign’s allomorphs throughout the family. Rule 16b partially conflicts with and must supersede Rule 8, and must follow Rules 1a and 13. It may also need to precede Rule 17.[21]
aw → aw, o·, a· (morphologically conditioned)

  • *we-3” + *aʔt- “place, put (TI)” + *-awINAN.OBJ(CL2)” + *-n(ay)N.FMV” + *-iINANsg” → *otaʔtni “s/he places it” (e.g., Ojibwe odatoon “s/he places it down; bets it”)
  • *aʔt- “place, put (TI)” + *-aw + *-ẅ3” + *-akiPROXpl” → *aʔtwaki “they place (something)” (e.g., Plains Cree ‡astāwak “they place it/them (INAN) there”)
  • *kwečiht- “try (TI)” + *-aw + *-(o)ko2pl.IMPER” → *kočihtawoko “try (pl.) it!” (e.g., Massachusett <qutchehteóꝏk> = †/kʷətʃəhtjawuːk/)
  • *weL-iht- “form (TI)” + *-aw + *-k3AN.CONJ” [← *-t by Rule 13] + *-e·SBJV” → *oL-ihtkw “if s/he forms it” (e.g., Unami ‡wəli·tá·we “if s/he makes/fixes it”)

(17) Word-final consonants or consonant clusters are lost. Rule 17 must follow Rule 16a and possibly Rule 16b.
(C)C(G) → Ø /__#

  • *ke-2” + *pankihšin- “fall (AI)” + *-eʔmFMV” + *-naw1pl.INCL” → **kepankihšineʔmenaw*kepankihšineʔmenaØ “we (INCL) fall” (e.g., Meskwaki ‡kepakishinepena)
  • *eškwa·nte·m- “door” [noun stem; to appear as a full noun needing one of the gender/number suffixes *-i or *-ari] → *eškwa·nte·Ø (adverbial particle) “at the door” (e.g., Munsee əskwáande, Ojibwe [gib]ishkwaand “beside the door,” Meskwaki [kep]ishkwâte “in the doorway”)

(18) *t and are palatalized to and respectively before *i(·) or *y. Rule 18 could be seen as always operational.
t, θ → č, š /__{i,i·,y}

  • *ne-1” + *-sit- “foot” + *-iINANsg” → *nesiči “my foot” (e.g., Shawnee niθiči); compare *ne- + *-sit- + *-ariINANpl” → *nesitari “my feet” (e.g., Shawnee niθitali)
  • *na·θ- “fetch (TA)” + *-i1OBJ” + *-an2sg.CONJ” → *na·šiyane· “if you fetch me” (e.g., Ojibwe naazhiyan); cf. *na·θ- + *-eθ2OBJ” + *-a·n1sg.CONJ” → *na·θeθa·ne· “if I fetch you” (e.g., Ojibwe naaninaan)

(19) There were various other minor rules that applied to individual morphemes in specific contexts (e.g., the 1pl.EXCL suffix *-na·n loses its final *n before some mode suffixes, the AI/II Final *-iwi “be” becomes *-ewi after stems ending in semivowels, and the combination of the irrealis *-w and third-person conjunct *-t [→ *-k by Rule 13] metathesizes to *kw [see example under Rule 13; and note the parallel with part of Rule 16b]), or vestiges of older rules which were no longer synchronically active in PA (e.g., the shift of *t to *s before *e and sometimes other vowels, in specific morphemes) which I have not catalogued here. Many can be found in Pentland (1979a, ch. 7).


Rules 9 as well as 4 and 16b above concerned the contraction of VGV sequences, but as noted these—especially the underlying sequence *|awe|—have proven difficult to reconstruct in Proto-Algonquian, because various daughter languages often show evidence of different kinds of contraction, or of no contraction at all. And in some cases the same sequence can be shown to have had different contraction outcomes in the protolanguage alone: compare, for example, the varying outcomes of the verb stem *no·ntaw- “hear (TA)” below: *ke-no·ntaw-ekw-a*keno·ntkwa “s/he hears you” vs. *ke-no·ntaw-eθ-eʔm*keno·ntθe “I hear you.” In spite of these complications, in this section I’ll give the best current understanding of the outcome of various VGV sequences in the PA period as described primarily by Goddard (2001b), and using mostly his examples from that paper. But be aware that some parts are still uncertain.


The underlying sequence *|awe| has three possible outcomes: uncontracted *awe, contraction to *o·, and contraction to *a·.

*aw-e contracts to *a· in verbal inflection before the inverse theme sign *-ekw:[22]

  • *ke-2” + *no·ntaw- “hear (TA)” + *-ekwINV” → *keno·ntkwa “s/he hears you” (e.g., Ojibwe ginoondaag, Meskwaki ‡kenôtâkwa)

*aw-e contracts to *o· in verbal inflection before the 2OBJ theme sign *-eθ, and in nominal inflection (e.g., before the locative suffix *-enki):

  • *ke-2” + *no·ntaw- “hear (TA)” + *-eθ2OBJ” → *keno·ntθe “I hear you” (e.g., Ojibwe ginoondoon, Meskwaki kenôtône “I heard what you said”)
  • *re·kaw- “sand” + *-enkiLOC” → *re·knki “on the sand” (e.g., Munsee léekoong)
  • *ne-1” + -i·yaw “body” + -(e)na·n “1pl.EXCL” → *ni·yna·ni “our (EXCL) bodies; ourselves (EXCL)” (e.g., Miami-Illinois niiyoonaani)

In most other cases, contraction evidently did not occur, including in verbal inflection before the morphemes *-em “marked obviative object” and *-ehko2pl»3AN.IMPER,” and in most derivation:

  • *ne-1” + *meθkaw- “find (TA)” + *-emMARK.OBV.OBJ” + *-a·3OBJ” + *-wFMV” + *-ariOBVsg” → *nemeθkawema·wari “I find h/ (OBV)” (e.g., Ojibwe nimikawimaan “I find him/her/them (OBV)”)
  • *maškaw- “hard, strong” + *-en+-am “by hand (TI(1))” → *maškawenamwa “s/he holds (something) tight, keeps a strong grip on (something)” (e.g., Ojibwe mashkawinam “s/he holds it tight, keeps a strong grip on it”)
  • *a·θaw- “fail” + *-esi “have property, be in state or condition (AI)” → *a·θawesiwa “s/he fails” (e.g., Meskwaki ânawesiwa)

Many of these patterns are disturbed in individual daughter languages. For example, Menominee and Delaware show contraction to *o· in verbal inflectional contexts where other languages reflect no contraction (e.g., Menominee hkon “(you pl.) use h/!” ← *aw-ehko); and while many languages reflect the contraction to *o· in nominal inflection, this is reversed in some daughters (including Ojibwe, which has, e.g., niiyawinaan “our (EXCL) bodies,” cf. the Miami-Illinois form). Still others, in these or other cases, have “recontraction” in which the archaic contraction pattern is replaced by a new one, normally one where the resulting vowel is a long version of the first vowel of the underlying VGV sequence—i.e., in this case, as though from *a· (e.g., Meskwaki nîyânâni “ourselves (EXCL)” ← *ni·yna·ni ← *|ne-i·yaw-(e)na·n-i|; cf. the Miami-Illinois and Ojibwe forms.)


*ewe, which primarily occured when the person prefixes were added to a stem beginning in *|we-|, contracted to *o·. This contraction is most obvious in Eastern Algonquian languages, because the other languages mostly came to treat non-dependent *o-initial stems like other vowel-initial stems and thus insert an epenthetic *-t- by Rule 2b, although the length on the /oː/ when preceded by a prefix usually remains. However, other vestiges of this contraction can occasionally be found outside the Eastern languages, as in the latter two examples below:

  • *ne-1” + *weL- “good; properly, arranged” + *-esi “have property, be in state or condition (AI)” → *nLesi “I am good [in appearance, etc.]” (e.g., Munsee nóolsi “I am pretty”)
  • *we- _ -i “possess (AI)” + *wexpwa·kan- “pipe” → **wewexpwa·kaniwa*xpwa·kaniwa “s/he has a pipe” (e.g., Plains Cree ōspwākaniw)
  • *ne-1” + *wekima·w- “chief” + *-emPOSSESSED.NOUN” → *nkima·ma “my chief” (e.g., older Plains Cree nōkimām, beside restructured nitōkimām with ni-t-)


In most cases, a *Vːwe sequence contracted to *Vː. This occurred in both nominal and verbal inflection and derivation:

  • *ne-1” + *na·pe·w- “man, male” + *-emPOSSESSED.NOUN” → *nena·pma “my husband” (e.g., Plains Cree nināpēm)
  • *eškote·w- “fire” + *-ehke· “obtain, make, use (AI)” → *eškothke·wa “s/he makes a fire” (e.g., Menominee esko·tε·hkεw, Meskwaki ashkotêhkêwa)
  • *maya·w- “directly” + *-eškaw “by foot/body (TA)” → *mayškaw- “meet head on (TA)” (e.g., Ojibwe omayaashkawaan “s/he (PROX) meets him/her/them (OBV) head on,” Meskwaki mayâshkawêwa “s/he (PROX) encounters/kicks h/ (OBV) squarely”)
  • *-ina·w “land of the X” + *-enkiLOC” → *-innki “in the land of/among the X” (e.g., Meskwaki ashâhinâki “in Sioux country, among the Sioux” ← Ashâha “Sioux person”)

However, there are traces of alternative outcomes. At least two verbs whose stems ended in *-e·w contracted to *o· instead of *e·: *ne·w- “see (TA)” and *wi·čye·w- “have along, accompany (TA)”:

  • *ne-1” + *ne·w- “see (TA)” + *-ekwINV” → *nenkwa “s/he sees me” (Shawnee ninookwa)

This contraction has been lost in all daughter languages besides Shawnee except for some fossilized, and thus clearly old, forms in Cree-Innu. In the other daughter languages these *-e·w verbs either contract to *e·, in line with the regular contraction pattern for -Vːw verb stems, or show no contraction, thus regularizing the surface stem shape: cf. Plains Cree niwīcēk “s/he has me along” (← *ne-1” + *wi·čye·w- “have along, accompany (TA)” + *-ekwINV”) and cognate Ojibwe niwiijiiwig “s/he accompanies me.”

There is also evidence for an old contraction of *a·we to *o· in some forms, though this aberrant pattern has been repaired in various ways by most of the daughters:

  • *nyi·pa·- “at night” + *-wehθe· “walk (AI)” → *nyi·phθe·wa “s/he walks in the dark” (Menominee ni·phnεw)

Because *wo· was an illegal sequence in PA after a consonant or word boundary, or when morpheme-internal, this *a·we*o· contraction seems not to have applied in pre-PA if *w preceded; in such cases the outcome was apparently *a· instead (Goddard 2007:259, n. 65):

  • *wi·ki- “dwell (AI)” + *-wa·w2pl/3pl” + *-eʔmNMLZ” → *wi·kiwʔmi “lodge, wigwam” (originally literally “(where) they dwell”; cf. Dakotan thípi “dwelling, lodge, tipi” with the same literal meanining).

Finally, it is likely that the sequence *a·we that arose in verbal inflection when the 3OBJ theme sign *-a· was followed by the W-formative *-w and one of the central suffixes *-(e)na·n or *-(e)naw (*-a·-w-ena·n / *-a·-w-enaw), or was followed by the irrealis suffix *-w and the UNSPEC»3AN.CONJ ending *-ent (*-a·-w-ent), did not contract, although Menominee does show later contraction to /oː/, several other languages show contraction of the former to *a· (which it’s possible was the real PA situation), and Meskwaki in the latter case shows contraction to /aː/ (Goddard 2007:236-237, 266, n. 73):

  • *ke-2” + *ne·w- “see (TA)” + *-a·3OBJ” + *-wFMV” + *-(e)naw1pl.INCL” → *kene·wa·wenawa “we (INCL) see h/” (e.g., Munsee kəneewáawəna and Massachusett <nunnauooun> †nəna·wą·wən “we [EXCL] saw him” [Goddard and Bragdon 1988:519; with prefix nə-1” in place of kə-2”]; cf., with contraction, Menominee kenε·wnaw and Eastern Mahican <N’wawehána> †nwa·wihą·nah “we (EXCL) know h/” [Goddard 2008:296; verb †wāwih-, prefix n-1”])
  • *neʔr- “kill (TA)” + *-a·3OBJ” + *-wIRR” + *-entUNSPEC»3AN.CONJ” + *-e·nINTERR” → *neʔra·wente·ni “whether s/he was killed” (e.g., Ojibwe nesaawinden [with initial change]; cf. Meskwaki nesâtêni “s/he must have been killed” with contraction)


*aye originally contracted to *e·, which is reflected in its value in Miami-Illinois and Eastern Algonquian. In most cases in the other languages it recontracted to *a·, except in Menominee and Shawnee, which restructured their nouns in *-ay, resulting in Menominee appearing to have recontracted to /iː/. That *e· was the original value is shown by relic/fossilized forms in languages which otherwise point to *a·:

  • *ne-1” + *či·pay- “dead person” + *-emPOSSESSED.NOUN” → *neči·pma “my corpse; my deceased relative” (e.g., Miami-Illinois ninciipeema “my deceased relative”; cf. Meskwaki nechîpâma with recontraction)
  • *apinay- “group resting place” + *-enkiLOC” → *apinnki “on the group resting place” (e.g., Munsee ăpíineeng “on the bed,” Miami-Illinois (a)pineenki “in bed”; cf. Meskwaki apinâki “on the family sitting place” with recontraction)
  • *ahθay- “skin, hide” + *-enkiLOC” → *ahθnki “on a skin/hide” (→ recontracted *ahθnki → Meskwaki asâki). For this same stem, cf. the fossilized form attesting to the original contraction pattern in *ahθay- + *-ehke· “obtain, make, use (AI)” → *ahθhke·wa “s/he tans hides” (Meskwaki asêhkêwa; and cf. cognate Ojibwe aseke, where the corresponding noun has been lost from the modern language)


In at least some cases, nouns in *-ay contracted with a following *i-initial derivational suffix to give *e·:

  • *či·pay- “dead person” + *-ina·w “land of the X” → *či·pna·wi “Land of the Dead” (e.g., Cheyenne áno = |séánó| “in the Land of the Dead” with *-enkiLOC”; Meskwaki chîpênâwesiwa “s/he is bothered by a ghost” with *-esi “have property, be in state or condition (AI)” [i.e., “*≈be among the dead”]; Ojibwe jiibenaake “s/he holds a Feast of the Dead” with *-ehke· “obtain, make, use (AI)” [i.e., “*≈get the dead”]) (Goddard 2001b:217)
  • *či·pay- + *-iθkanaw “path” → *či·pθkanawi “the Path of the Dead” (e.g., Ojibwe jiibekana “Path of the Dead; the Milky Way” [the souls of the dead are believed to travel along the Milky Way]; cf. the more recent regularized variant jiibay-miikana with miikana “road, path”)

However, verbs of possession (with the AI Final *-i) apparently contracted *ayi to *a·:

  • *we- _ -i “possess (AI)” + *-htawakay- “ear” → **ohtawakayiwa*ohtawakwa “s/he has ears” (e.g., Ojibwe otawagaa, Plains Cree ohtawakāw)

Other *VGe

Most other instances of *VG-e contracted to *Vː (Pentland 1999:251). Goddard discusses several examples. Noun stems in *-iy (which are actually always stems in *-Cwiy), contract *iye to *i· in both inflection and derivation:

  • *asa·twiy- “poplar (Populus spp.)” + *-enkiLOC” → *asa·twnki “at the poplar(s)” (e.g., Ojibwe azaadiing, Arapaho hohóótiiʔ “at the tree(s)”)
  • *apanšwiy- “lodgepole” + *-ehke· “obtain, make, use (AI)” → *apanšwhke·wa “s/he makes/gathers lodgepoles” (e.g., Kickapoo apasiihkea)

Likewise, *iw-e and *ow-e contract to *i· and *o· respectively before a consonant:

  • *erenyiw- “man” + *-ewi “be (AI/II)” → *erenywiwa “he is a man” (e.g., Ojibwe ininiiwi, Arapaho hiineníínit, Cheyenne [é]hetanéveo’o = |-etanéve-o| (pl.) “they are men,” Swampy Cree ininīwiw “s/he is alive”)
  • *we-3” + *mi·či- “eat (TI)” + *-hsiDIM” + *-wIRR” + *-(e)n(ay)FMV” + *-ariINANpl” → *omi·čihsnari “s/he doesn’t eat them (INAN) a little bit” (e.g., Ojibwe [gaawiin] omiijisiinan “s/he doesn’t eat them (INAN)”)
  • *ke-2” + *men- “drink (AI)” + *-hsDIM” + *-ow [← *-w by Rule 6] “IRR” + *-eʔmFMV” + *-naw1pl.INCL” → *kemensʔmena “we (INCL) don’t drink a little bit” (Miami-Illinois ‡kimensoomina “we (INCL) don’t drink,” Cheyenne né[sáa]manéhema = |né=sáa=mane-ʹhé-ma| “we (INCL) don’t drink”)
  • *pye·siw- “bring (hither) (TA)” + *-ehko2pl»3AN.IMPER” → *pye·shko “(you pl.) bring him/her/them!” (e.g., Plains Cree pēsīhko[k] “(you pl.) bring them!”)

Cheyenne, Menominee, Massachusett, and the Delaware languages show a lack of contraction, at least in some forms, of noun stems in *-iw, but these seem to be later analogical developments. These same languages also show a lack of contraction in corresponding forms in the inflection of other noun stems in *-Vw, such as PA *eškote·w- “fire” (compare the discussion of *Vːwe contraction above).

  • *wačiw- “mountain” + *-enkiLOC” → *wačnki “at/on the mountain” (e.g., Ojibwe wajiing) → later decontracted post-PA *wačiwenki (e.g., Cheyenne vȯsēva = |vosev-á|, synchronically |vose-vá|)
  • *eškote·w- “fire” + *-enki*eškotnki “at/in the fire” (e.g., Ojibwe ishkodeng) → later decontracted post-PA *eškote·wenki (e.g., Menominee esku·tiah = |ɛskotɛːw-Eh|, cf. contracted Menominee esko·tε·hkεw “s/he makes a fire” above)
  • *ne-1” + *-eθkwe·w- “woman” + *-emPOSSESSED.NOUN” → *neteθkwma “my woman, my sister (m.s.)” (e.g., Ojibwe nindikwem “my wife, my woman”) → later decontracted post-PA *neteθkwe·wema (e.g., Unami ntuxkwé·yum “my sister (m.s.)” [← pre-Unami *nətəxkwé·wəm])


A large number of Proto-Algonquian words has been reconstructed, including two dictionaries, Aubin (1975) and Hewson (1993), but there are major caveats. The basic vocabulary, especially basic nouns such as names for a number of animals and trees, features of the natural world, basic tools and other domestic objects, etc., is uncontroversial, and some of these were monomorphemic. Interestingly, a quite significant number, especially animal and plant names, were polymorphemic, including *e·hsepana “raccoon” (“shell-eater” ?), *ka·kwa “porcupine” (“prickly bushy-tailed mammal”), *šeka·kwa “skunk” (“pissing bushy-tailed mammal”), *wa·poswa “hare” (“white (game?) animal”), and the majority of tree and bush names (which usually end in *-a·xkw- “wood; hardwood/deciduous tree,” *-a·ntakw- “evergreen tree,” or *-eminšy- “tree, bush; fruit/nut-bearing plant”) and fruit/berry/grain names (which end in *-min-, as in *ote·himini “strawberry,” lit. “heart berry”) (Siebert 1967b:21, 25-29, 36; Pentland 1979a:332-335, 372). There were still plenty of names for animals, tools, and elements of the natural world which are unanalyzable, however. A very significant number of bird names were reduplicated (Cowan 1972), but otherwise mostly monomorphemic.

Virtually no Proto-Algonquian verb stems were monomorphemic, due to the way Algonquian languages build words, and especially verbs, out of Initials, Medials, and Finals (see here for a brief overview of this in Ojibwe). While we can theoretically reconstruct thousands of verbs back to PA—and this has indeed been done—in many cases these may reflect daughter languages independently combining the same shared morphological elements.

While as noted two PA dictionaries exist, both are far from ideal. Aubin’s is now out of date and was assembled from a number of disparate sources dating back to the 1920s, with little attempt to normalize or update anything or to avoid duplication of entries; it is, in a sense, more of a bibliographical protolanguage dictionary than an etymological one. If just used for this purpose, at least, it’s quite useful. Hewson’s dictionary was mechanically generated by a computer comparing lexemes in Bloomfield’s original four languages, with basically no recourse to other languages, so it is geographically and genetically biased and very frequently unreliable. Other than these two dictionaries, most Proto-Algonquian reconstructions are scattered throughout dozens of individual papers and books. The late David Pentland’s magnum opus, a full, massive etymological dictionary for the family, is in preparation and will hopefully be published in the near future.


There is general agreement on the Proto-Algonquian numerals, especially those from “one” to “five,” which don’t pose much difficulty in reconstruction, although most Western languages have restructured many of their numerals for “six” through “ten” by suffixing *tahθwi “(be) so many (in number),” usually with subsequent phonetic reduction. Additionally, only some of the lower numerals have clear cognates in Blackfoot, so many of these are more strictly speaking Proto-Core Algonquian reconstructions. The numerals are given in the table below (Rhodes and Costa 2003; with Wiyot forms from Teeter 1964b:91-92, and Yurok forms from Garrett 2014:106-107 and Robins 1958:86-89, and the Yurok forms somewhat abstracted by necessity). Some patterning should be apparent: each of the Proto-Core Algonquian numbers from “one” through “five” ended in a stem-final *-w and began with *n(y)-, while “six,” “seven,” and “eight” are based on “one,” “two,” and “three” with an added suffix *-a·ši(ka) “plus five.”

  PA (cf. Blackfoot) (cf. Wiyot/Yurok)
one *nekotwi; *pe·šekwi ni’t- W. kuts-; Y. koht-
two *nyi·šwi naat-/niist- W. ḍit-; Y. nVʔ-
three *neʔθwi (ni(iw)- (cog.?)) W. ḍikh-; Y. nahkṣ-
four *nye·wi (niis- (cog.?)) W. ḍiyaʔw-; (Y. toʔon-)
five *nya·θanwi nisit- (W. weʔsàɣ); (Y. meruh-)
six *nekotwa·ši(ka) (náa-) (W. təkłəlúk); Y. kohč[ew]
seven *nyi·šwa·ši(ka) (ihkitsiki-) (W. háʔləw); (Y. čṛwṛṣik’)
eight *neʔθwa·ši(ka) (naanisi-) (W. híwitəw); (Y. knewetik’)
nine *ša·nka (piihkssi-) (W. βəšəṛúk); (Y. kṛ:mik’)
ten *metaθa (kiip-) (W. ḍəlúk); (Y. wełowa:)


Unfortunately, little work has been done on reconstructing Proto-Algonquian syntax above the level of the word, although more work has been done on comparative Algonquian syntax and the syntax of individual Algonquian languages, especially in the last couple of decades. As a broad generalization of cross-Algonquian syntax we can note a few points, which I will expand on in a future post. Algonquian languages are polysynthetic, very strongly head-marking, and sometimes described as non-configurational, though the last point is controversial and several researchers have argued against such an analysis. Pan-Algonquian syntactic phenomena include some of the expected traits of highly polysynthetic and (at least superficially) non-configurational languages: pragmatically determined word order (see below), widespread null anaphora, discontinuous constituents, and noun incorporation, in addition to processes such as quantifier floating, and raising of complement verb subjects to main clause objects. Another feature found throughout Algonquian are “relative roots,” Initials, or sometimes verb stems, which add a semi-oblique element as an additional valence to a verb—but without making it a normal verbal argument (they are not true applicatives).

Bloomfield (1958:131) wrote of Ojibwe that “[w]ord order is decidedly flexible,” and it is true that no Algonquian language uses word order simply to mark grammatical relations in the way that languages like English do; however, many syntactic processes, including word order, are driven by pragmatic/discourse factors, and are not merely random. There is a complex interplay of grammatical status (subject vs. primary object vs. secondary object vs. oblique), animacy, obviation, “direction”/voice (direct vs. inverse, if relevant) of the verb, definiteness, givenness, topic, and focus in determining the placement of overt NPs with respect to the verb and to each other. Dahlstrom’s (1993, 1995) studies focusing on Meskwaki seem to have been particularly influential, and her analysis, envisioning a very flat phrase structure as typical of a non-configurational language, is often reproduced, even by researchers who have alternative analyses:

Proposed clause structure for Meskwaki
Proposed clause structure for Meskwaki: there is no VP, all elements except the verb are optional, the default position for non-verbal XPs is postverbal, and all fronted elements are daughters of S except for an element which occupies the Topic position, which is a daughter of Sʹ. (Based on Dahlstrom 1993:13 and 1995:3.)

An example from Meskwaki illustrating many of these elements is below, with Dahlstrom’s analysis; again, more detailed discussion on comparative Algonquian syntax is deferred until a future post.

MESKWAKI    (Dahlstrom 2017:63)
     Înokiwîna âkwi kêkôhi kehkênetakini mana mehtosêneniwa.
     înoki⸗wîna    âkwi  kêkôhi      kehkênetakini               [mana          mehtosêneniwa]
     today⸗CONTR  NEG     anything  s/he.doesn’  [this(PROX)  person                   ]
     TOPIC                NEG   FOCUS         V                                           XP (SUBJECT)
     “But today the people don’t know anything.”

Sources Used [click to expand]

(“AA” = American Anthropologist [new series])
(“AENA” = Archaeology of Eastern North America)
(“AIL” = Algonquian and Iroquoian Linguistics)
(“AIL-M” = Algonquian and Iroquoian Linguistics Memoir)
(“AL” = Anthropological Linguistics)
(“APS-M” = Memoirs of the American Philosophical Society)
(“CSLI” = Center for the Study of Language and Information)
(“CWPL” = Calgary Working Papers in Linguistics)
(“HNAI” = Handbook of North American Indians, series ed. William C. Sturtevant)
(“IJAL” = International Journal of American Linguistics)
(“JIPA” = Journal of the International Phonetic Association)
(“JLSM” = Janua Linguarum, Studia Memoriae Nicolai van Wijk Dedicata)
(“KWPL” = Kansas Working Papers in Linguistics)
(“SCOIL” = Survey of California and Other Indian Languages”)
(“TIL-SM” = Trends in Linguistics, Studies and Monographs)
(“UBCWPL” = University of British Columbia Working Papers in Linguistics)
(“UCPL” = University of California Publications in Linguistics)
(“WSCLA” = Proceedings of the Annual Workshop on the Structure and Constituency of the Languages of the Americas)

  • Ames, Kenneth M., Don E. Dumond, Jerry R. Galm, and Rick Minor (1998). “Prehistory of the Southern Plateau.” In Plateau, ed. Deward E. Walker, Jr., pp. 103-119. HNAI vol. 12. Washington, D.C.: Smithsonian Institution.
  • Aubin, George F. (1975). A Proto-Algonquian Dictionary. National Museum of Man, Mercury Series, Canadian Ethnology Service Paper 29. Ottawa: National Museums of Canada.
  • Béland, Jean Pierre (1978). Atikamekw Morphology and Lexicon. Ph.D. dissertation, University of California, Berkeley.
  • Berman, Howard (1992). “A Restriction on the Shape of Proto-Algonquian Nouns.” IJAL 58(3): 321-323. DOI: 10.1086/ijal.58.3.3519790.
  • Berman, Howard (2006). “Studies in Blackfoot Prehistory.” IJAL 72(2): 264-284. DOI: 10.1086/507167.
  • BkD = Frantz, Don[ald G.], and Inge Genee, eds. (2015–). Blackfoot Dictionary. Algonquian Dictionaries Project. Accessed October 29, 2020.
  • Bloomfield, Leonard (1925). “On the Sound-System of Central Algonquian.” Language 1(4): 130-156. DOI: 10.2307/409540.
  • Bloomfield, Leonard (1946). “Algonquian.” In Linguistic Structures of Native America by Harry Hoijer et al., ed. Cornelius Osgood, pp. 85-129. Viking Fund Publications in Anthropology 6. New York: Viking Fund. Online version prepared by Will Oxford available here.
  • Bloomfield, Leonard (1958). Eastern Ojibwa: Grammatical Sketch, Texts and Word List, ed. Charles F. Hockett. Ann Arbor: University of Michigan Press.
  • Bloomfield, Leonard (1962). The Menomini Language, ed. Charles F. Hockett. New Haven, CT: Yale University Press.
  • Brittain, Julie (2000). “A Metrical Analysis of Primary Stress Placement in Southern East Cree.” IJAL 66(2): 181-217. DOI: 10.1086/466417.
  • Brunton, Bill B. (1998). “Kootenai.” In Plateau, ed. Deward E. Walker, Jr., pp. 223-237. HNAI vol. 12. Washington, D.C.: Smithsonian Institution.
  • Campbell, Lyle (1997). American Indian Languages: The Historical Linguistics of Native America. Oxford Studies in Anthropological Linguistics 4. New York: Oxford University Press.
  • Catalogue of Endangered Languages (2019). University of Hawai’i at Manoa. Accessed November 27, 2019.
  • CDO = Fisher, Louise, Wayne Leman, Leroy Pine Sr., and Marie Sanchez (2017). Cheyenne Dictionary Online. Chief Dull Knife College. Accessed February 27, 2020.
  • Cenerini, Chantale, Marie-Odile Junker, and Nicole Rosen (2017). “Mapping Dialect Variation Using the Algonquian Linguistic Atlas.” Language Documentation & Conservation 11: 305-324.
  • Chatters, James C., and David L. Pokotylo (1998). “Prehistory: Introduction.” In Plateau, ed. Deward E. Walker, Jr., pp. 73-80. HNAI vol. 12. Washington, D.C.: Smithsonian Institution.
  • Costa, David J. (1992). “Miami-Illinois Animal Names.” AIL 17(3): 19-44.
  • Costa, David J. (1996). “Reconstructing Initial Change in Algonquian.” AL 38(1): 39-72.
  • Costa, David J. (2001). “Shawnee Noun Plurals.” AL 43(3): 255-287.
  • Costa, David J. (2003). The Miami-Illinois Language. Lincoln: University of Nebraska Press.
  • Costa, David J. (2007). “The Dialectology of Southern New England Algonquian.” In Papers of the Thirty-Eighth Algonquian Conference, ed. H.C. Wolfart, pp. 81-127. Winnipeg: University of Manitoba.
  • Costa, David J. (2013). “Borrowing in Southern Great Lakes Algonquian and the History of Potawatomi.” AL 55(3): 195-233. DOI: 10.1353/anl.2013.0013.
  • Costa, David J. (2021). “The Emphatic Negative in Indiana Miami and the Dialectology of Miami-Illinois.” In Thomason et al. (2021), pp. 63-81.
  • Cowan, William (1972). “Reduplicated Bird Names in Algonquian.” IJAL 38(4): 229-230. DOI: 10.1086/465220.
  • Cowan, William (1983). “The Development of Suprasegmental Inflections in Montagnais.” IJAL 49(1): 64-71. DOI: 10.1086/465765.
  • Cowell, Andrew (2020). “The Noowo3ineheeno’ Language.” Paper presented at the 52nd Algonquian Conference, October 23, 2020.
  • Cowell, Andrew (2021). “The Phonology of the Noowoθineheenoʔ Language, and Its Relation to Arapaho and Cheyenne.” Ms., based on Cowell (2020).
  • Dahlstrom, Amy (1986). Plains Cree Morphosyntax. Ph.D. dissertation, University of California, Berkeley.
  • Dahlstrom, Amy (1993). “The Syntax of Discourse Functions in Fox.” In Proceedings of the Nineteenth Annual Meeting of the Berkeley Linguistics Society, February 12-15, 1993: Special Session on Syntactic Issues in Native American Languages, ed. David A. Peterson, pp. 11-21. Berkeley: Berkeley Linguistics Society, University of California. DOI: 10.3765/bls.v19i2.1558.
  • Dahlstrom, Amy (1995). Topic, Focus, and Other Word Order Problems in Algonquian. The Belcourt Lecture, delivered before the University of Manitoba on 25 February 1994. Winnipeg: Voices of Rupert’s Land.
  • Dahlstrom, Amy (2013). “Argument Structure of Quirky Algonquian Verbs.” In From Quirky Case to Representing Space: Papers in Honor of Annie Zaenen, eds. Tracy Holloway King and Valeria de Paiva, pp. 61-71. Stanford, CA: CSLI.
  • Dahlstrom, Amy (2017). “Seeking Consensus on the Fundamentals of Algonquian Word Order.” In Papers of the Forty-Fifth Algonquian Conference, eds. Monica Macaulay, Margaret Noodin, and J. Randolph Valentine, pp. 59-72. East Lansing: Michigan State University Press. Preprint version available here.
  • DAtk = Chachai, Véronique, Marie-Odile Junker, and Nicole Petiquay, eds. (2015–). Dictionnaire Atikamekw. Algonquian Dictionaries Project. Accessed October 19, 2020.
  • DeBlois, A[lbert] D., ed. (1967). Contributions to Anthropology: Linguistics I (Algonquian). National Museum of Canada Bulletin 214. Ottawa: Queen’s Printer.
  • Dempsey, Hugh A. (2001). “Blackfoot.” In Plains, ed. Raymond J. DeMallie, pp. 604-628. HNAI vol. 13. Washington, D.C.: Smithsonian Institution.
  • Denny, J. Peter (1989). “Algonquian Connections to Salishan and Northeastern Archaeology.” In Actes du Vingtième Congrès des Algonquinistes, ed. William Cowan, pp. 86-107. Ottawa: Carleton University.
  • Denny, J. Peter (1991). “The Algonquian Migration from Plateau to Midwest: Linguistics and Archaeology.” In Papers of the Twenty-Second Algonquian Conference, ed. William Cowan, pp. 103-124. Ottawa: Carleton University.
  • Dryer, Matthew S. (2007). “Kutenai, Algonquian and the Pacific Northwest from an Areal Perspective.” In Papers of the Thirty-Eighth Algonquian Conference, ed. H.C. Wolfart, pp. 155-206. Winnipeg: University of Manitoba.
  • Ellis, C. Douglas (1971). “Cree Verb Paradigms.” IJAL 37(2): 76-95. DOI: 10.1086/465143.
  • Ellis, C. Douglas (2015). “Cree Verb Paradigms.” Hosted by the Algonquian Dictionaries Project. [Revised and updated version of paradigms in Ellis (1971).]
  • ESMCD = Ellis, C. Douglas, and Marie-Odile Junker, eds. (2018–). Eastern Swampy and Moose Cree Dictionary. Algonquian Dictionaries Project. Accessed June 20, 2019.
  • Fiedel, Stuart J. (1990). “Middle Woodland Algonquian Expansion: A Refined Model.” North American Archaeologist 11(3): 209-230. DOI: 10.2190/U241-WB50-PJT6-0HE4.
  • Fiedel, Stuart J. (1991). “Correlating Archaeology and Linguistics: The Algonquian Case.” Man in the Northeast 41: 9-32.
  • Fiedel, Stuart J. (2013). “Are Ancestors of Contact Period Ethnic Groups Recognizable in the Archaeological Record of the Early Late Woodland?” AENA 41: 221-229.
  • Fiedel, Stuart J. (2020). “The Origins of Northern Native Americans: New Insights and Enigmas from Ancient DNA.” Ms., ResearchGate. DOI: 10.13140/RG.2.2.29069.36326.
  • Foster, Michael K. (1996). “Language and the Culture History of North America.” In Languages, ed. Ives Goddard, pp. 64-110. HNAI vol. 17. Washington, D.C.: Smithsonian Institution.
  • Frantz, Donald G. (2009). Blackfoot Grammar. 2nd edition. Toronto: University of Toronto Press.
  • Galm, Jerry R. (1994). “Prehistoric Trade and Exchange in the Interior Plateau of Northwestern North America.” In Prehistoric Exchange Systems in North America, eds. Timothy G. Baugh and Jonathon E. Ericson, pp. 275-305. Interdisciplinary Contributions to Archaeology. New York: Springer. DOI: 10.1007/978-1-4757-6231-0_10.
  • Garrett, Andrew (2001). “Reduplication and Infixation in Yurok: Morphology, Semantics, and Diachrony.” IJAL 67(3): 264-312. DOI: 10.1086/466460.
  • Garrett, Andrew (2014). Basic Yurok. SCOIL Report 16. Berkeley: SCOIL, University of California, Berkeley.
  • Goddard, Ives (1966). “[Review of] The Wiyot Language, by Karl V. Teeter.” IJAL 32(4): 398-404. DOI: 10.1086/464930.
  • Goddard, Ives (1967a). “The Algonquian Independent Indicative.” In DeBlois (1967), pp. 66-106.
  • Goddard, Ives (1967b). “Corrections [to Goddard (1966)].” IJAL 33(3): 261. DOI: 10.1086/464975.
  • Goddard, Ives (1967c). “Notes on the Genetic Classification of the Algonquian Languages.” In DeBlois (1967), pp. 7-12.
  • Goddard, Ives (1973). “Proto-Algonquian *nl and *nθ.” IJAL 39(1): 1-6. DOI: 10.1086/465232.
  • Goddard, Ives (1974a). “An Outline of the Historical Phonology of Arapaho and Atsina.” IJAL 40(2): 102-116. DOI: 10.1086/465292.
  • Goddard, Ives (1974b). “Remarks on the Algonquian Independent Indicative.” IJAL 40(4): 317-327. DOI: 10.1086/465328.
  • Goddard, Ives (1975). “Algonquian, Wiyot, and Yurok: Proving a Distant Genetic Relationship.” In Linguistics and Anthropology in Honor of C. F. Voegelin, eds. M. Dale Kinkade, Kenneth L. Hale, and Oswald Werner, pp. 249-262. Lisse: Peter de Ridder Press.
  • Goddard, Ives (1978a). “Central Algonquian Languages.” In Northeast, ed. Bruce G. Trigger, pp. 583-587. HNAI vol. 15. Washington, D.C.: Smithsonian Institution.
  • Goddard, Ives (1978b). “Eastern Algonquian Languages.” In Northeast, ed. Bruce G. Trigger, pp. 70-77. HNAI vol. 15. Washington, D.C.: Smithsonian Institution.
  • Goddard, Ives (1979a). “Comparative Algonquian.” In The Languages of Native America: Historical and Comparative Assessment, eds. Lyle Campbell and Marianne Mithun, pp. 70-132. Austin: University of Texas Press.
  • Goddard, Ives (1979b). Delaware Verbal Morphology: A Descriptive and Comparative Study. Outstanding Dissertations in Linguistics. New York: Garland.
  • Goddard, Ives (1980). “Eastern Algonquian as a Genetic Subgroup.” In Papers of the Eleventh Algonquian Conference, ed. William Cowan, pp. 143-158. Ottawa: Carleton University.
  • Goddard, Ives (1981a). “Against the Linguistic Evidence Claimed for Some Algonquian Dialectal Relationships.” AL 23(7): 271-297.
  • Goddard, Ives (1981b). “Massachusett Phonology: A Preliminary Look.” In Papers of the Twelfth Algonquian Conference, ed. William Cowan, pp. 57-105. Ottawa: Carleton University.
  • Goddard, Ives (1982). “The Historical Phonology of Munsee.” IJAL 48(1): 16-48. DOI: 10.1086/465711.
  • Goddard, Ives (1983). “The Eastern Algonquian Subordinative Mode and the Importance of Morphology.” IJAL 49(4): 351-387. DOI: 10.1086/465800.
  • Goddard, Ives (1987). “Leonard Bloomfield’s Descriptive and Comparative Studies of Algonquian.” Historiographia Linguistica 14(1/2): 179-217. DOI: 10.1075/hl.14.1-2.17god.
  • Goddard, Ives (1988). “Pre-Cheyenne *y.” In Shipley (1988), pp. 345-360. DOI: 10.1515/9783110852387.345.
  • Goddard, Ives (1990a). “Algonquian Linguistic Change and Reconstruction.” In Linguistic Change and Reconstruction Methodology, ed. Philip Baldi, pp. 99-114. TIL-SM 45. Berlin: Mouton de Gruyter. DOI: 10.1515/9783110886092.99.
  • Goddard, Ives (1990b). “Primary and Secondary Stem Derivation in Algonquian.” IJAL 56(4): 449-483. DOI: 10.1086/466171.
  • Goddard, Ives (1991). “Observations Regarding Fox (Mesquakie) Phonology.” In Papers of the Twenty-Second Algonquian Conference, ed. William Cowan, pp. 157-181. Ottawa: Carleton University.
  • Goddard, Ives (1994a). Leonard Bloomfield’s Fox Lexicon: Critical Edition. AIL-M 12. Winnipeg: Algonquian and Iroquoian Linguistics.
  • Goddard, Ives (1994b). “The West-to-East Cline in Algonquian Dialectology.” In Actes du Vingt-Cinquième Congrès des Algonquinistes, ed. William Cowan, pp. 187-211. Ottawa: Carleton University.
  • Goddard, Ives (1996a). “Introduction.” In Languages, ed. Ives Goddard, pp. 1-16. HNAI vol. 17. Washington, D.C.: Smithsonian Institution.
  • Goddard, Ives, compiler (1996b). “Native Languages and Language Families of North America.” Map insert. In Languages, ed. Ives Goddard. HNAI vol. 17. Washington, D.C.: Smithsonian Institution.
  • Goddard, Ives (2000). “The Historical Origin of Cheyenne Inflections.” In Papers of the Thirty-First Algonquian Conference, ed. John D. Nichols, pp. 77-129. Winnipeg: University of Manitoba.
  • Goddard, Ives (2001a). “The Algonquian Languages of the Plains.” In Plains, ed. Raymond J. DeMallie, pp. 71-79. HNAI vol. 13. Washington, D.C.: Smithsonian Institution.
  • Goddard, Ives (2001b). “Contraction in Fox (Meskwaki).” In Actes du Trente-Deuxième Congrès des Algonquinistes, ed. John D. Nichols, pp. 164-230. Winnipeg: University of Manitoba.
  • Goddard, Ives (2001c). “The Languages of the Plains: Introduction.” In Plains, ed. Raymond J. DeMallie, pp. 61-70. HNAI vol. 13. Washington, D.C.: Smithsonian Institution.
  • Goddard, Ives (2002a). “Explaining the Double Reflexes of Word-Initial High Short Vowels in Fox.” Diachronica 19(1): 43-80. DOI: 10.1075/dia.19.1.03god.
  • Goddard, Ives (2002b). “Grammatical Gender in Algonquian.” In Papers of the Thirty-Third Algonquian Conference, ed. H.C. Wolfart, pp. 195-231. Winnipeg: University of Manitoba.
  • Goddard, Ives (2003). “Reconstructing the History of the Demonstrative Pronouns of Algonquian.” In Rudes and Costa (2003), pp. 37-102.
  • Goddard, Ives (2004). “Meskwaki Verbal Affixes.” In Papers of the Thirty-Fifth Algonquian Conference, ed. H.C. Wolfart, pp. 97-123. Winnipeg: University of Manitoba.
  • Goddard, Ives (2006). “The Proto-Algonquian Negative and Its Descendants.” In Actes du Trente-Septième Congrès des Algonquinistes, ed. H.C. Wolfart, pp. 161-208. Winnipeg: University of Manitoba.
  • Goddard, Ives (2007). “Reconstruction and History of the Independent Indicative.” In Papers of the Thirty-Eighth Algonquian Conference, ed. H.C. Wolfart, pp. 207-271. Winnipeg: University of Manitoba.
  • Goddard, Ives (2008). “Notes on Mahican: Dialects, Sources, Phonemes, Enclitics, and Analogies.” In Papers of the Thirty-Ninth Algonquian Conference, eds. Karl S. Hele and Regna Darnell, pp. 246-315. London, ON: University of Western Ontario.
  • Goddard, Ives (2010). “Linguistic Variation in a Small Speech Community: The Personal Dialects of Moraviantown Delaware.” AL 52(1): 1-48.
  • Goddard, Ives (2013). “The Munsee of Charles Halfmoon’s Translations.” In Papers of the Forty-First Algonquian Conference, eds. Karl S. Hele and J. Randolph Valentine, pp. 81-119. Albany: State University of New York Press.
  • Goddard, Ives (2015a). “Arapaho Etymologies.” Ms.
  • Goddard, Ives (2015b). “Arapaho Historical Morphology.” AL 57(4): 345-411. DOI: 10.1353/anl.2016.0010.
  • Goddard, Ives (2015c). “Three Nineteenth-Century Munsee Texts: Archaisms, Dialect Variation, and Problems of Textual Criticism.” In New Voices for Old Words: Algonquian Oral Literatures, ed. David J. Costa, pp. 198-314. Lincoln: University of Nebraska Press.
  • Goddard, Ives (2016). “The ‘Loup’ Languages of Western Massachusetts: The Dialectal Diversity of Southern New England Algonquian.” In Papers of the Forty-Fourth Algonquian Conference, eds. Monica Macaulay, Margaret Noodin, and J. Randolph Valentine, pp. 104-138. Albany: State University of New York Press.
  • Goddard, Ives (2018). “Blackfoot and Core Algonquian Inflectional Morphology: Archaisms and Innovations.” In Papers of the Forty-Seventh Algonquian Conference, eds. Monica Macaulay and Margaret Noodin, pp. 83-106. East Lansing: Michigan State University Press.
  • Goddard, Ives (2020). “The Paradox of Partial Similarity: Unami Secondary-Object Marking and Other Clones of the N-Endings.” In Papers of the Forty-Ninth Algonquian Conference, eds Monica Macaulay and Margaret Noodin, pp. 99-113. East Lansing: Michigan State University Press.
  • Goddard, Ives, and Kathleen J. Bragdon (1988). Native Writings in Massachusett. 2 volumes. APS-M 185. Philadelphia: American Philosophical Society.
  • Goddard, Ives, and Lucy Thomason (2014). A Meskwaki-English and English-Meskwaki Dictionary, Based on Early Twentieth-Century Writings by Native Speakers. Petoskey, MI: Mundart Press.
  • Golla, Victor, ed. (1984). The Sapir-Kroeber Correspondence: Letters Between Edward Sapir and A. L. Kroeber, 1905-1925. SCOIL Report 6. Berkeley: SCOIL, University of California, Berkeley.
  • Golla, Victor (2007). “North America.” In Encyclopedia of the World’s Endangered Languages, ed. Christopher [J.] Moseley, pp. 1-95. Abingdon, UK: Routledge. Book DOI: 10.4324/9780203645659.
  • Golla, Victor (2011). California Indian Languages. Berkeley: University of California Press.
  • Golla, Victor, Ives Goddard, Lyle Campbell, Marianne Mithun, and Mauricio Mixco (2007). “North America.” In Atlas of the World’s Languages, eds. R.E. Asher and Christopher J. Moseley, pp. 7-41. 2nd edition. Abingdon, UK: Routledge. Book DOI: 10.4324/9781315829845.
  • Haas, Mary R. (1958). “Algonkian-Ritwan: The End of a Controversy.” IJAL 24(3): 159-173. DOI: 10.1086/464453.
  • Harvey, Chris (2005). Sound Change in Old Montagnais. MA thesis, University of Manitoba.
  • Hasler, Laurel Anne, Marie-Odile Junker, Yvette Mollen, and Claire Owen (2020). “Pedagogical and User Guide.” Innu Lessons. Accessed January 21, 2021.
  • Hewson, John (1973). “Proto-Algonkian Reflexes in Micmac.” AL 15(3): 151-164.
  • Hewson, John (1983). “Some Micmac Etymologies.” In Actes du Quatorzième Congrès des Algonquinistes, ed. William Cowan, pp. 301-306. Ottawa: Carleton University.
  • Hewson, John (1993). A Computer-Generated Dictionary of Proto-Algonquian. Hull, QC: Canadian Museum of Civilization. Online version with minor corrections available here.
  • Hewson, John (n.d.). Proto Algonkian Roots and Proto Algonkian Word Formatives. E-Book, Department of Linguistics, Memorial University of Newfoundland.
  • Hockett, C[harles] F. (1981a). “The Phonological History of Menominee.” AL 23(2): 51-87.
  • Hockett, C[harles] F. (1981b). “Errata in ‘The Phonological History of Menominee.’” AL 23(7): 326.
  • Howell, Benita J., Richard S. Levy, and Alvin Luckenbach (1979). “What Is Dalrymple’s Pamunkey?” IJAL 45(1): 78-80. DOI: 10.1086/465576.
  • Julian, Charles (2010). A History of the Iroquoian Languages. Ph.D. dissertation, University of Manitoba.
  • Junker, Marie-Odile, Mimie Neacappo, Claire Owen, and Laurel Anne Hasler (2019). “User and Pedagogical Guide.” East Cree Lessons. Accessed January 21, 2021.
  • Keller, Christine K. (2009). Glacial Kame Sandal-Sole Shell Gorgets: An Exploration of Manufacture, Use, Distribution, and Public Exhibition. MA thesis, Ball State University.
  • Kinkade, M. Dale, William W. Elmendorf, Bruce Rigsby, and Haruo Aoki (1998). “Languages.” In Plateau, ed. Deward E. Walker, Jr., pp. 49-72. HNAI vol. 12. Washington, D.C.: Smithsonian Institution.
  • Koontz, John E. (1990). “Avian Loans: Selected Birdnames of Eastern North America.” Ms., February 13, 1990 revision of paper presented at the American Anthropological Association, November 20, 1988.
  • Lacombe, A[lbert] (1874). Grammaire de la Langue des Cris. Montréal: Beauchemin & Valois.
  • LeSourd, Philip S. (1988). Accent and Syllable Structure in Passamaquoddy. Ph.D. dissertation, Massachusetts Institute of Technology.
  • LeSourd, Philip S. (2003a). “The Noun Substitute in Maliseet-Passamaquoddy.” In Rudes and Costa (2003), pp. 141-163. Preprint version available here.
  • LeSourd, Philip S. (2003b). “Traces of Proto-Algonquian *wi·la ‘he, she’ in Maliseet-Passamaquoddy.” IJAL 69(4): 357-370. DOI: 10.1086/382737.
  • LTD = Rementer, Jim, Bruce Pearson, Jan Brown, Nicky Kay Michael, April Daniels, and Ray Whritenour, eds. (2002–). Lenape Talking Dictionary. Accessed January 30, 2022.
  • MacKenzie, Marguerite (1980). Towards a Dialectology of Cree-Montagnais-Naskapi. Ph.D. dissertation, University of Toronto.
  • MacKenzie, Marguerite (2003). “Cree Dialects of Canada.” Map. Accessed January 1, 2021.
  • Madsen, Carl D. (1997). “Microwear Analysis of the Lithic Assemblage at the Rosenberger Site.” Journal of California and Great Basin Anthropology 19(1): 116-123.
  • MCOD = Brousseau, Kevin, ed. (2014–). Moose Cree Online Dictionary. Moose Factory’s Community Language Project, Moose Cree First Nation. Accessed September 28, 2022.
  • Meillet, A[ntoine] (1925). La Méthod Comparative en Linguistique Historique. Instituttet for Sammenlingende Kulturforsking, Serie A: Forelesninger, II. Oslo: H. Aschehoug & Co.
  • Michelson, Truman (1914). “Two Alleged Algonquian Languages of California.” AA 16(2): 361-367. DOI: 10.1525/aa.1914.16.2.02a00150.
  • Michelson, Truman (1915). “Rejoinder [to Sapir (1915a)].” AA 17(1): 194-198. DOI: 10.1525/aa.1915.17.1.02a00270.
  • Michelson, Truman (1935). “Phonetic Shifts in Algonquian Languages.” IJAL 8(3/4): 131-171. DOI: 10.1086/463813.
  • Miller, Wick R. (1996). “The Ethnography of Speaking.” In Languages, ed. Ives Goddard, pp. 222-243. HNAI vol. 17. Washington, D.C.: Smithsonian Institution.
  • Milligan, Marianne (2005). Menominee Prosodic Structure. Ph.D. dissertation, University of Wisconsin, Madison.
  • Mithun, Marianne (1984). “The Proto-Iroquoians: Cultural Reconstruction from Lexical Materials.” In Extending the Rafters: Interdisciplinary Approaches to Iroquoian Studies, eds. Jack Campisi, Michael K. Foster, and Marianne Mithun, pp. 259-282. Albany: State University of New York Press.
  • Mithun, Marianne (1999). The Languages of Native North America. Cambridge Language Surveys. Cambridge: Cambridge University Press.
  • Mithun, Marianne (2007). “Integrating Approaches to Diversity: Argument Structure on the NW Coast.” In Diversity in Language: Perspectives and Implications, eds. Yoshiko Matsumoto, David Y. Oshima, Orrin R. Robinson, and Peter Sells, pp. 9-36. CSLI Lecture Notes 176. Stanford: CSLI. Preprint version available here.
  • Mithun, Marianne (2012). “Core Argument Patterns and Deep Genetic Relations: Hierarchical Systems in Northern California.” In Argument Structure and Grammatical Relations: A Crosslinguistic Typology, eds. Pirkko Suihkonen, Bernard Comrie, and Valery Solovyev, pp. 257-294. Studies in Language Companion Series 126. Amsterdam: John Benjamins. DOI: 10.1075/slcs.126.12mit.
  • Moctezuma Zamarrón, José Luis (2011). El Sistema Fonológico del Kickapoo de Coahuila, Analizado desde las Metodologías Distribucional y Funcional. [Mexico City], Mexico: Instituto Nacional de Lenguas Indígenas.
  • Mooney, James (1907). “The Powhatan Confederacy, Past and Present.” AA 9(1): 129-152.
  • Muehlbauer, Jeffrey (2012). “Vowel Spaces in Plains Cree.” JIPA 42(1): 91-105. DOI: 10.1017/S0025100311000302.
  • Nichols, Johanna (1971). “Diminutive Consonant Symbolism in Western North America.” Language 47(4): 826-848. DOI: 10.2307/412159.
  • Nilsen, Campbell (2017). The Historical Development of the Menominee Independent Indicative. Unpublished senior thesis.
  • Nilsen, Campbell (2019). “Revisiting the Historical Context of Some Menominee Morphophonological Rules.” In Papers of the Forty-Eighth Algonquian Conference, eds. Monica Macaulay and Margaret Noodin, pp. 191-203. East Lansing: Michigan State University Press.
  • O’Meara, John (1990). Delaware Stem Morphology. Ph.D. dissertation, McGill University.
  • OPD = Nichols, John D., and Nora Livesay, eds. (2012–). Ojibwe People’s Dictionary. Department of American Indian Studies, University of Minnesota. Accessed September 27, 2022.
  • Oxford, Will (2007). Towards a Grammar of Innu-aimun Particles. MA thesis, Memorial University of Newfoundland.
  • Oxford, Will (2014). Microparameters of Agreement: A Diachronic Perspective on Algonquian Verb Inflection. Ph.D. dissertation, University of Toronto.
  • Oxford, Will (2015). “Patterns of Contrast in Phonological Change: Evidence from Algonquian Vowel Systems.” Language 91(2): 308-357. DOI: 10.1353/lan.2015.0028.
  • Oxford, Will (2016). “Proto-Algonquian Phonotactics.” Ms.
  • Oxford, Will (2019). “Fission in Algonquian and the Status of Morphological Templates.” In WSCLA 23, eds. D. K. E. Reisinger and Roger Yu-Hsiang Lo. Vancouver: UBCWPL.
  • Oxford, Will (2020a). “Algonquian.” In The Routledge Handbook of North American Languages, eds. Daniel Siddiqi, Michael Barrie, Carrie Gillon, Jason D. Haugen, and Éric Mathieu, pp. 504-523. Routledge Handbooks in Linguistics. New York: Routledge. DOI: 10.4324/9781315210636-21. Preprint version available here.
  • Oxford, Will (2020b). “Direct, Inverse, and Neutral: Refining the Description of Algonquian Transitive Verb Forms.” Paper presented at the 52nd Algonquian Conference, October 25, 2020.
  • Oxford, Will (2021). “The Meaning of -w.” Paper presented at the 53rd Algonquian Conference, October 16, 2021.
  • Oxford, Will (forthcoming). The Algonquian Inverse. Oxford University Press.
  • Parks, Douglas R., and Robert L. Rankin (2001). “Siouan Languages.” In Plains, ed. Raymond J. DeMallie, pp. 94-114. HNAI vol. 13. Washington, D.C.: Smithsonian Institution.
  • PbD = Penobscot Dictionary (2015). Penobscot Cultural & Historic Preservation. Accessed September 27, 2022.
  • PCD = Wolvengrey, Arok, ed. (2015–). Plains Cree Dictionary. Algonquian Dictionaries Project. Accessed June 20, 2019.
  • Pentland, David H. (1974). “Diminutive Consonant Symbolism in Algonquian.” In Papers of the 6th Algonquian Conference, ed. William Cowan, pp. 237-252. National Museum of Man Mercury Series, Canadian Ethnology Service Paper 23. Ottawa: National Museums of Canada.
  • Pentland, David H. (1977). “Proto-Algonquian *št.” IJAL 43(3): 225-226. DOI: 10.1086/465484.
  • Pentland, David H. (1979a). Algonquian Historical Phonology. Ph.D. dissertation, University of Toronto.
  • Pentland, David H. (1979b). “Causes of Rapid Phonological Change: The Case of Atsina and Its Relatives.” CWPL 5: 99-138. DOI: 10.11575/PRISM/29006.
  • Pentland, David H. (1979c). “Does ‘Eastern Algonquian’ Really Exist?” Algonquian Linguistics 4(4): 36-41.
  • Pentland, David H. (1983). “Proto-Algonquian [č] and [š].” In Actes du Quatorzième Congrès des Algonquinistes, ed. William Cowan, pp. 379-396. Ottawa: Carleton University.
  • Pentland, David H. (1984). “New Modes in Old Ojibwa.” AIL 9: 11-17.
  • Pentland, David H. (1991). “Mahican Historical Phonology.” In Schmick’s Mahican Dictionary by Carl Masthay, pp. 15-27. APS-M 197. Philadelphia: American Philosophical Society.
  • Pentland, David H. (1992). “Palatalization in Narragansett.” Handout for paper presented at the 24th Algonquian Conference, October 24, 1992.
  • Pentland, David H. (1996). “The Southern Algonquians and Their Neighbours.” In Papers of the Twenty-Seventh Algonquian Conference, ed. David H. Pentland, pp. 224-237. Winnipeg: University of Manitoba.
  • Pentland, David H. (1998). “Initial *S > N in Arapaho-Atsina.” Diachronica 15(2): 309-321. DOI: 10.1075/dia.15.2.05pen.
  • Pentland, David H. (1999). “Morphology of the Algonquian Independent Order.” In Papers of the Thirtieth Algonquian Conference, ed. David H. Pentland, pp. 222-266. Winnipeg: University of Manitoba.
  • Pentland, David H. (2000a). “Inaccessible and Absentative Inflections in Algonquian.” AIL 25(3): 25-26.
  • Pentland, David H. (2000b). “The Origins of the Cree Demonstrative Pronouns.” AIL 25(4): 35-38.
  • Pentland, David H. (2003). “The Missinipi Dialect of Cree.” In Papers of the Thirty-Fourth Algonquian Conference, ed. H.C. Wolfart, pp. 287-302. Winnipeg: University of Manitoba.
  • Pentland, D[avid] H. (2006). “Algonquian and Ritwan Languages.” In Encyclopedia of Language & Linguistics. 14 volumes. 2nd edition, ed. Keith Brown, vol. 1, pp. 161-166. Oxford: Elsevier. DOI: 10.1016/B0-08-044854-2/02265-3.
  • Pentland, David H. (2013). “Nominal Inflection in Nipmuck.” In Papers of the Forty-First Algonquian Conference, eds. Karl S. Hele and J. Randolph Valentine, pp. 230-267. Albany: State University of New York Press.
  • Pleger, Thomas C. (2000). “Old Copper and Red Ocher Social Complexity.” Midcontinental Journal of Archaeology 25(2): 169-190. DOI: 10.2307/20708133.
  • PMD = Francis, David A., Robert M. Leavitt, and Margaret Apt (2008–). Passamaquoddy-Maliseet Dictionary. Language Keepers. Accessed September 27, 2022.
  • Proulx, Paul (1977). “Connective Vowels in Proto-Algonquian.” IJAL 43(2): 156-157. DOI: 10.1086/465473.
  • Proulx, Paul (1980a). “The Linguistic Evidence on Algonquian Prehistory.” AL 22(1): 1-21.
  • Proulx, Paul (1980b). “The Subordinative Order of Proto-Algonquian.” IJAL 46(4): 289-300. DOI: 10.1086/465664.
  • Proulx, Paul (1982). “The Origin of the Absolute Verbs of the Algonquian Independent Order.” IJAL 48(4): 394-411. DOI: 10.1086/465749.
  • Proulx, Paul (1984a). “Algonquian Objective Verbs.” IJAL 50(4): 403-423. DOI: 10.1086/465850.
  • Proulx, Paul (1984b). “Proto-Algic I: Phonological Sketch.” IJAL 50(2): 165-207. DOI: 10.1086/465826.
  • Proulx, Paul (1984c). “Proto-Algonquian *aye and Its Implications.” IJAL 50(1): 84-93. DOI: 10.1086/465816.
  • Proulx, Paul (1984d). “Two Models of Algonquian Linguistic Prehistory: Diffusion Versus Genetic Subgrouping.” AL 26(4): 393-434.
  • Proulx, Paul (1985). “Proto-Algic II: Verbs.” IJAL 51(1): 59-93. DOI: 10.1086/465860.
  • Proulx, Paul (1988). “The Demonstrative Pronouns of Proto-Algonquian.” IJAL 54(3): 309-330. DOI: 10.1086/466088.
  • Proulx, Paul (1989). “A Sketch of Blackfoot Historical Phonology.” IJAL 55(1): 43-82. DOI: 10.1086/466104.
  • Proulx, Paul (1990). “Proto-Algonquian Verb Inflection.” KWPL 15(2): 100-145. DOI: 10.17161/KWPL.1808.440.
  • Proulx, Paul (2001). “Proto-Algonquian Demonstratives and the Size of PA Social Units.” AIL 26(2): 15-19. [Possible preprint version, taken from Proulx (2005b), pp. 53-65.]
  • Proulx, Paul (2003a). “The Evidence on Algonquian Genetic Grouping: A Matter of Relative Chronology.” AL 45(2): 201-225.
  • Proulx, Paul (2003b). “[Review of] Actes du Trente-Deuxième Congrès des Algonquinistes, Edited by John D. Nichols.” IJAL 69(4): 409-414. DOI: 10.1086/382741.
  • Proulx, Paul (2004a). “Malecite-Passamaquoddy wìl and Algonquian Subgrouping.” UniverSOS 1: 165-174.
  • Proulx, Paul (2004b). “[Review of] Blair A. Rudes and David J. Costa (eds.), Essays in Algonquian, Catawban, and Siouan Linguistics in Memory of Frank T. Siebert, Jr.Anthropologica 46(1): 114-116. DOI: 10.2307/25606183.
  • Proulx, Paul (2005a). “Initial Change in Blackfoot.” CWPL 26: 1-26. DOI: 10.11575/PRISM/28977.
  • Proulx, Paul (2005b). Reviews and Commentary (1983-2004). Lulu Press.
  • Proulx, Paul (2005c). “[Review of] Blair A. Rudes and David J. Costa (eds.), Essays in Algonquian, Catawban, and Siouan Linguistics in Memory of Frank T. Siebert, Jr.” Preprint version of Proulx (2004b) contained in Proulx (2005b), pp. 18-25.
  • Proulx, Paul (n.d.). “Central Algonquian as a Genetic Grouping, and Dialect Area.” Ms., downloaded from author’s website in 2005 but no longer accessible.
  • Rhodes, Richard A. (2020). “Language Shift in the Subarctic and Central Plains.” In The Language of Hunter-Gatherers, eds. Tom Güldemann, Patrick McConvell, and Richard A. Rhodes, pp. 552-576. Cambridge: Cambridge University Press. DOI: 10.1017/9781139026208.021.
  • Rhodes, Richard A. (2021). “The Case for Core Central Algonquian.” In Thomason et al. (2021), pp. 305-345.
  • Rhodes, Richard A. (n.d.). “On the Classification of Central Algonquian.” Ms., based on a paper presented at the 20th Algonquian Conference, October 29, 1988.
  • Rhodes, Richard A., and David J. Costa (2003). “The History of Algonquian Number Words.” In Rudes and Costa (2003), pp. 181-215.
  • Rhodes, Richard A., and Evelyn M. Todd (1981). “Subarctic Algonquian Languages.” In Subarctic, ed. June Helm, pp. 52-66. HNAI vol. 6. Washington, D.C.: Smithsonian Institution.
  • Rigsby, Bruce, and Noel Rude (1996). “Sketch of Sahaptin, a Sahaptian Language.” In Languages, ed. Ives Goddard, pp. 666-692. HNAI vol. 17. Washington, D.C.: Smithsonian Institution.
  • Robins, R. H. (1958). The Yurok Language: Grammar, Texts, Lexicon. UCPL 15. Berkeley: University of California Press.
  • Roll, Tom E., and Steven Hackenberger (1998). “Prehistory of the Eastern Plateau.” In Plateau, ed. Deward E. Walker, Jr., pp. 120-137. HNAI vol. 12. Washington, D.C.: Smithsonian Institution.
  • Rudes, Blair A., and David J. Costa, eds. (2003). Essays in Algonquian, Catawban and Siouan Linguistics in Memory of Frank T. Siebert, Jr. AIL-M 16. Winnipeg: Algonquian and Iroquoian Linguistics.
  • Sapir, Edward (1913). “Wiyot and Yurok, Algonkin Languages of California.” AA 15(4): 617-646. DOI: 10.1525/aa.1913.15.4.02a00040.
  • Sapir, Edward (1915a). “Algonkin Languages of California: A Reply [to Michelson (1914)].” AA 17(1): 188-194. DOI: 10.1525/aa.1915.17.1.02a00270.
  • Sapir, Edward (1915b). “Epilogue [Reply to Michelson (1915)].” AA 17(1): 198. DOI: 10.1525/aa.1915.17.1.02a00270.
  • Shipley, William, ed. (1988). In Honor of Mary Haas: From the Haas Festival Conference on Native American Linguistics. Berlin: Mouton de Gruyter. DOI: 10.1515/9783110852387.
  • Siebert, Frank T., Jr. (1941). “Certain Proto-Algonquian Consonant Clusters.” Language 17(4): 298-303. DOI: 10.2307/409280.
  • Siebert, Frank T., Jr. (1967a). “Discrepant Consonant Clusters Ending in *-k in Proto-Algonquian: A Proposed Interpretation of Saltatory Sound Changes.” In DeBlois (1967), pp. 48-59.
  • Siebert, Frank T., Jr. (1967b). “The Original Home of the Proto-Algonquian People.” In DeBlois (1967), pp. 13-47.
  • Siebert, Frank T., Jr. (1975). “Resurrecting Virginia Algonquian from the Dead: The Reconstituted and Historical Phonology of Powhatan.” In Studies in Southeastern Indian Languages, ed. James M. Crawford, pp. 285-453. Athens: University of Georgia Press.
  • Snow, Dean R. (1975). “The Archaeological Implications of the Proto-Algonquian Urheimat.” In Papers of the Seventh Algonquian Conference, ed. William Cowan, pp. 339-346. Ottawa: Carleton University.
  • Snow, Dean R. (1996). “More on Migration in Prehistory: Accommodating New Evidence in the Northern Iroquoian Case.” American Antiquity 61(4): 791-796. DOI: 10.2307/282019.
  • Starna, William A. (2017). “After the Handbook: A Perspective on 40 Years of Scholarship Since the Publication of the Handbook of North American Indians, Volume 15, Northeast.” New York History 98(1): 112-146. DOI: 10.1353/nyh.2017.0040.
  • Strachey, William (1953 [1612]). The Historie of Travell into Virginia Britania, eds. Louis B. Wright and Virginia Freund. London: Hakluyt Society.
  • [Symonds, William =] “W. S.” (1612). A Map of Virginia, with a Description of the Countrey, the Commodities, People, Government and Religion . . . , ed. Thomas Abbay. Oxford: Joseph Barnes.
  • Taylor, Allan R. (1960). “Blackfoot Historical Phonology: A Preliminary Survey.” With handwritten annotations by Mary R. Haas. Ms., item #Haas.068.003 in the collection Miscellaneous Papers from the Survey of California and Other Indian Languages. SCOIL, University of California, Berkeley. Accessed September 5, 2020.
  • Teeter, Karl V. (1964a). “Algonquian Languages and Genetic Relationship.” In Proceedings of the Ninth International Congress of Linguists, Cambridge, Mass., August 27-31, 1962, ed. Horace G. Lunt, pp. 1026-1034. JLSM Series Maior 12. The Hague: Mouton.
  • Teeter, Karl V. (1964b). The Wiyot Language. UCPL 37. Berkeley: University of California Press.
  • Thomason, L[ucy G.] (2006). “Proto-Algonkian Phonology and Morpho-Syntax.” In Encyclopedia of Language & Linguistics. 14 volumes. 2nd edition, ed. Keith Brown, vol. 10, pp. 190-198. Oxford: Elsevier. DOI: 10.1016/B0-08-044854-2/01896-4.
  • Thomason, Lucy, David J. Costa, and Amy Dahlstrom, eds. (2021). Webs of Relationships and Words from Long Ago: A Festschrift Presented to Ives Goddard on the Occasion of His 80th Birthday. Petoskey, MI: Mundart Press.
  • Thompson, Laurence C., and M. Dale Kinkade (1990). “Languages.” In Northwest Coast, ed. Wayne Suttles, pp. 30-51. HNAI vol. 7. Washington, D.C.: Smithsonian Institution.
  • Walker, Deward E., Jr., and Roderick Sprague (1998). “History Until 1846.” In Plateau, ed. Deward E. Walker, Jr., pp. 138-148. HNAI vol. 12. Washington, D.C.: Smithsonian Institution.
  • Warne, Janet (1975). A Historical Phonology of Abenaki. MA thesis, McGill University.
  • Weber, Natalie (2016a). “Accent and Prosody in Blackfoot Verbs.” In Papers of the Forty-Fourth Algonquian Conference, eds. Monica Macaulay, Margaret Noodin, and J. Randolph Valentine, pp. 348-369. Albany: State University of New York Press.
  • Weber, Natalie (2016b). “Blackfoot Reflexes of Proto-Algonquian Clusters in *ʔ and *h.” Handout for paper presented at the 48th Algonquian Conference, October 13-16, 2016.
  • Weber, Natalie (2016c). “Initial Extrametricality and Cyclicity in Blackfoot Accent.” In UBC Qualifying Papers 3 (2013-2015), eds. Andrei Anghelescu, Joel Dunham, and Natalie Weber, pp. 234-249. UBCWPL 44.
  • Weber, Natalie (2017). “Blackfoot Reflexes of Proto-Algonquian Clusters.” Handout for paper presented at the 23rd International Conference on Historical Linguistics, July 23-31, 2017.
  • Weber, Natalie, and Jason A. Shaw (2022). “Situating Blackfoot Within a Typology of (Mobile) Boundary Tone Grammars.” In Proceedings of the 2021 Annual Meeting on Phonology, eds. Peter Jurgec, Liisa Duncan, Emily Elfner, Yoonjung Kang, Alexei Kochetov, Brittney K. O’Neill, Avery Ozburn, Keren Rice, Nathan Sanders, Jessamyn Schertz, Nate Shaftoe, and Lisa Sullivan. Washington, D.C.: Linguistic Society of America. DOI: 10.3765/amp.v9i0.5154.
  • WLD = Butler, Lynnika, Herman Leung, Andrew Garrett, and Elizabeth Lieu, eds. (2014–). Wiyot Language Database. Wiyot Tribe / Department of Linguistics, University of California, Berkeley. Accessed March 23, 2020.
  • Wolfart, H. Christoph (1973). Plains Cree: A Grammatical Study. Transactions of the American Philosophical Society 63, part 5. Philadelphia: American Philosophical Society. DOI: 10.2307/1006246.

  1. Extinct or dormant languages and dialects, those with no living fluent native speakers, are marked with “†”. Languages or dialects which we know both existed and were Algonquian but for which we have no or essentially no actual certain records are included in some cases, but marked with “(?)”: Ha’anahawunena, said to be the most divergent of the “Arapahoan” group of languages; Mascouten, part of the Meskwakian group; the Sutaio dialect of Cheyenne; and the Wappinger dialect of Munsee. Etchemin, which is only securely attested in a single list of numerals, is also marked with “(?)” since its classification is uncertain beyond the fact that it’s Eastern Algonquian and was evidently spoken in Maine. (This should not be confused with Maliseet-Passamaquoddy, for a long period referred to by some as “Etchemin,” which was the name Samuel de Champlain gave the Maliseet people he first encountered at Tadoussac.) On the complicated situation involving the SNEA languages and particularly those traditionally labeled “Loup A” and “Loup B,” see footnote three.

    Although plenty of Algonquianists would disagree on some details, and in particular on whether some groupings shown here represent genetic subgroups versus areal ones, this probably reasonably represents the approximate current consensus, such as there is. A few additional subgroups which have been suggested at one point or another have not been included here: Cheyenne-Menominee, Southeastern Algonquian (Nanticoke-Conoy + Virginia Algonquian + Carolina Algonquian), SNEA-Southeastern, and Delawaran-Wabanakian. In the second and third cases we are hampered by the meager and poor-quality records of all of the Southeastern languages and most of the Southern New England languages; and I see nothing especially suggestive that the Southeastern languages form any coherent grouping. (There are a few intriguing links between SNEA and Nanticoke in particular, though.) In the fourth case, I have only seen Pentland (2005:243, n. 15) mention such a group, without giving the evidence (beyond a neutralization of length in high vowels supposedly not shared by other Eastern languages).

    One final grouping not shown is one recently proposed by Rhodes (2021), “Common East Algonquian,” which would encompass Eastern Algonquian + Core Central. The evidence Rhodes offers does not support such a subgroup; see my discussion of the question here. I’m also not completely convinced that Core Central itself is even a genetic rather than an areal grouping. Core Central in turn is often divided into two subgroups, Northern Core Central (Miami-Illinois and Ojibwe-Potawatomi) and Southern Core Central (Meskwakian and Shawnee). Southern Core Central looks very likely to me, even if the rest of Core Central doesn’t form a subgroup, but I’m doubtful whether Miami-Illinois and Ojibwe-Potawatomi are any closer to one another genetically than either is to the Southern Core Central languages. The few unique traits they do share seem superficial.

    In the early 1980s, and to some extent afterwards, the classification of Algonquian languages was the subject of heated debate between Paul Proulx and Ives Goddard (Proulx 1980a, 1980b, 1984a, 1984c, 1984d, 2003a; Goddard 1981a, 1983). Specifically, based among other things on his reconstruction of the history of the “subordinative mode” of verbs which is found in Eastern Algonquian languages, along with other considerations, Proulx denied the existence of an Eastern Algonquian subgroup (which had been argued for by Goddard 1967a, 1967c, 1979a, and 1980 inter alia), instead proposing, over time, a series of various alternative groupings. In the latest version he held before his death (e.g., Proulx 2003a, 2004a:166-168, n.d.), he continued to deny Eastern Algonquian, instead claiming that all non-Eastern languages formed a “Central Algonquian” subgroup, that the Core Central languages formed a subgroup (his “Lake” languages) within “Central Algonquian,” and that the three smaller subgroups of Eastern Algonquian languages proposed by Siebert (1975) (SNEA-Southeastern; Delawaran; and Abenakian) were also valid genetic units. Proulx failed to convince other Algonquianists, and some form of Eastern Algonquian as a genetic subgroup is now accepted by everyone, including those who were initially skeptical. Separately from this whole brouhaha, over time there has developed an increasing exploration and acceptance of the existence of lower-level subgroups both within and outside Eastern Algonquian.

    Finally, there is of course not always a clear line between “divergent dialects” and “closely related languages,” and to a good extent the breakdown of Cree-Innu, Ojibwe, and AGV into dialects vs. languages is semi-arbitrary, and in the first two cases overly conservative. There is a very strong break in intelligibility between western varieties of Cree on the one hand and East Cree, Innu, and Naskapi on the other, hence my treating them as two separate languages, but within both western Cree and the eastern Cree-Innu group there are numerous varities that are unintelligible to speakers of other varities within the same broader group, and some other fairly sharp intelligibility breaks (e.g., one roughly between Western and Eastern Swampy Cree, and one between Atikamekw and the rest of western Cree). The many varieties of East Cree, Innu, and Naskapi diverge from one another significantly, which is masked by a common Innu orthography. Realistically it is more accurate to consider the Cree-Innu varities as a small language family that contains some dialect continua, rather analogous to the situation among the Germanic varieties which stretch from Central Europe to the North Sea. Ojibwe could similarly easily be considered to consist of a number of not-always-quite-mutually-intelligible languages that nevertheless form dialect continua in many areas. Meanwhile, Gros Ventre and Arapaho-Besawunena could be considered divergent dialects of a single language.

    One final note: it has long been customary to refer to various Algonquian languages, especially Cree-Innu varieties and the languages and dialects of Southern New England, by what reflex they show for PA *r (and ) (e.g., as “r-dialects” and “n-dialects” and so on), in a way that implies these reflexes have some significance for subgrouping. But while this may sometimes be useful, I feel it is often misleading, and I have avoided using such terminology here, except obliquely in some later discussion of dialectology.

  2. The conventional wisdom classifies Nawathinehena and the undocumented Ha’anahawunena as “Arapahoan,” with Arapaho-Gros Ventre then forming a subgroup within Arapahoan as a sister clade to Nawathinehena. (Ha’anahawunena was said to have been the most divergent of the five “Arapahoan” varieties, supposedly even in some way “intermediate” between Arapaho and Blackfoot, but since it’s undocumented we can’t confirm this.) Nawathinehena does share a number of phonological developments with AGV, but there are also a number of significant differences (Cowell 2020, 2021; and cf. Pentland 1979a:179-200 with Goddard 1974a). Crucially, many of the shared developments are ones which took place relatively late or are also shared with the clearly non-Arapahoan language Cheyenne, or differ in certain details from the equivalent change in AGV, while some of the significant differences involve changes that must have occurred relatively early in the history of AGV and Nawathinehena (as first noted by Goddard 1974a:105).

    To me, all this suggests that Nawathinehena most likely did not form a genetic subgroup with AGV, but rather that after a period of independent development its speakers came in contact with AGV (and later Cheyenne) speakers and shared in many of the areally diffused distinctive phonological developments of both languages, though clearly sharing far more with AGV. (Cowell 2020, 2021:[16-19], and Pentland 1979a:176-177 come to the same conclusion.) But the picture is complicated by the fact that it appears to have been very similar to Arapaho and Gros Ventre morphologically (Cowell 2020, 2021). In any case, because of its uncertain status I have not classified it as Arapahoan here. Unfortunately our only documentation of Nawathinehena is a set of 145 words of less than perfect quality, so a definitive answer to the question may never be possible. Since Ha’anahawunena is entirely undocumented, I have also not classified it as Arapahoan.

  3. Sorting out the Algonquian languages originally spoken in southern New England and their relationships with one another is no easy task, both because most are very sparsely attested, and because a large number of the books or manuscripts which supposedly contain material on a given language can in fact be shown to contain material from multiple languages, or at least multiple linguistic varieties. In this footnote I’ll try to explain some of the labels used.

    First, it seems evident that Massachusett-Narragansett, spoken in eastern mainland Massachusetts, the islands of Nantucket and Martha’s Vineyard, and much of Rhode Island, constituted a dialect continuum. Massachusett, in its multiple dialects, is the best attested SNEA language, with translations of the Bible and other religious materials, vocabularies, a grammar, and numerous documents written by native speakers such as wills, land deeds, and petitions. The modern reconstituted version of Massachusett is referred to as Wampanoag (or Wôpanâak, in the practical orthography), which nowadays is often the name applied to the older language too, but I think it’s useful to be able to distinguish the two.

    Narragansett is supposedly the language described in Roger Williams’s 1643 book A Key into the Language of America, but in fact the majority of that book can be shown to represent a variety other than Narragansett Proper—probably the Coweset dialect of Massachusett, spoken in central Rhode Island. (Politically, the Cowesets were a subgroup of the Narragansetts.) Some of the forms in Williams’s book, however, do match what we would expect Narragansett to look like based on statements made by Williams and others, and so can probably be identified as Narragansett Proper. There are also traces of a third and probably even a fourth variety in Williams’s book; the identity of these languages is not known, though one is probably one of the “Loup A” languages (see below). Meanwhile, the “Narragansett” in Ezra Stiles’s short 1769 vocabulary doesn’t match Williams’s materials and also doesn’t match what is expected of genuine Narragansett, and is most likely actually a record of Eastern Niantic, which was spoken just to the south of Narragansett, on the southern tip of Rhode Island, and was evidently a dialect of Mohegan-Pequot. The Eastern Niantics and Narragansetts had merged after King Philip’s War, and the combined group was called the “Narragansetts,” but “the Eastern Niantics appear to have been numerically dominant,” and it was evidently their speech which was used by the whole group by the 1760s (Costa 2007:113).

    Second, there are two documents from the 18th century containing materials from Algonquian languages for which we have no other name than “Loup,” the French word for “wolf,” which they applied to numerous Algonquian-speaking groups in New England. The varieties represented in these documents have conventionally been labeled “Loup A” (the longer, 124-page document, compiled by Fr. Jean-Claude Mathevet) and “Loup B” (a shorter 14-page document by an unknown author; traditionally attributed to Fr. François-August Magon de Terlaye, but while it occurs inside a larger manuscript of his it does not seem to be in his handwriting [Pentland 2013:265; Goddard 2016:105, n. 2, citing p.c. from Pentland]). Most of the Loup A manuscript appears to represent a single language that is clearly closely related to Massachusett-Narragansett, and has been determined on various grounds to most likely represent the language of the Nipmucks of central Massachusetts. However, it also contains fragments of at least two other varieties, which Goddard (2007, 2008, 2016) labels “Loup 2” and “Loup 3” (with the main/Nipmuck portion of Loup A being “Loup 1”). The Loup B manuscript also represents a mixture of multiple dialects or languages, which Goddard labels “Loup 4” (according to him, apparently the same language as “Loup 3” of the Loup A manuscript), “Loup 5,” and “Loup 6.” The different Loup varieties show varying degrees of shared features with (other) SNEA languages, with Mahican, and/or with Western Abenaki. Goddard (2016) uses this to attempt to tease out their approximate geographical locations, and thus the most likely identity of their speakers. His conclusion is that Loup 1 = Nipmuck, Loup 2 = Norwottuck and/or Agawam, Loup 3/4 = Pocumtuck, Loup 5 = possibly Woronoco, and Loup 6 = possibly Pojassick; all of these groups except the Nipmucks were members of the Pocumtuck Confederacy and located in the Connecticut Valley in western Massachusetts.

    (Note that Loup 6, at least, was not an SNEA language, though even most of the other Loup varieties show some features that are out of step with “canonical” SNEA languages. I would in any case caution that other than the fact that both of these manuscripts contain material from at least three varieties—and probably the identification of at least one of the Loup A varieties with Nipmuck—which we can be fairly confident about, much of this is quite speculative, and at some points rather circular. I should also note that Pentland [2013] identifies Loup 2 as the “majority dialect”—and thus “the ‘real’ Nipmuck”—represented in the Loup A manuscript, not Loup 1, though he views all of the Loup A materials as different dialects of Nipmuck.)

    Third, Quiripi and Naugatuck were apparently dialects of a single language, but our materials on them are far from ideal. Quiripi is known from a single poorly translated catechism, and once again the catechism in fact clearly reflects two different varieties; the one with PEA *r reflected as /r/ is probably “real” Quiripi, while the one which has /j/ for *r may have been the language of the Wangunks (aka Mattabesecs) of central Connecticut (Costa 2007:116-119; Goddard 2016:133). What is generally called Naugatuck is essentially only attested in one very short wordlist. Rudes (1997) gives “Paugussett” and “Potatuck” as additionally attested varieties of Quiripi-Naugatuck, although the attestations are incredibly meager; he also treats the vocabulary others refer to as “Naugatuck” as being in the Quiripi dialect.

    Rudes (1997) treats Unquachog (central Long Island), which is almost entirely attested in a single word list recorded by Thomas Jefferson, as a dialect of Quiripi-Naugatuck, though others sometimes treat it as a separate, closely related language. Once again, the “Unquachog” vocabulary contains a mixture of forms from two different Algonquian varieties; Goddard (2016:133) suggests the non-“real” Unquachog portion represents the speech of the Matinecocks and Massapequas of western Long Island, while Costa (2007:121) suggests that the non-Unquachog portion is from a variety to the east and that the Matinecocks and Massapequas spoke Munsee.

    Note finally that two different linguistic entities have been called “Wampano” at one time or another: a long-extinct dialect of Munsee on the one hand, and Quripi-Naugatuck-Unquachog on the other. I will use “Wampano” to refer to the Munsee dialect. Along similar lines, the Mohegan language and Mohegan people should not be confused with the Mahican language/people, whose name is also spelled “Mohican.” Though he was not the first to confuse them, James Fenimore Cooper’s conflation of the two in his fictionalized “Mohicans” (based primarily on the Mahicans, but using the name of a famous Mohegan chief for the character of Uncas) has certainly helped perpetuate such confusion into the present.

  4. In this post, Wiyot and Yurok will be written using variants of their traditional technical orthographies: Wiyot /i, u, ə, ɛ, ɑ/ = <i, u, ə, e, a>, Yurok /i, u, ɚ, ɛ, ɔ, a/ = <i, u, ṛ, e, o, a>, and long vowels indicated with a following colon <:>. Remaining characters match their International Phonetic Alphabet values, with the following exceptions: <y>, <ṣ>, <š>, <č>, and <c> are used in place of [j], [ʂ], [ʃ], [tʃ], and [ts], and the retroflex tap/flap [ɽ] and approximant [ɻ] are written <ḍ> and <ṛ> (except that in Yurok I write consonantal [ɻ] with <r>).

    Wiyot has traditionally been written with <a> for /ə/ and <o> for /ɑ/; aspiration and labialization marked by a following <h>/<w> rather than a superscript <h>/<w>; <b>, <d>, and <g> for /β/, /ɽ/, and /ɣ/; <r> for /ɻ/; and (sometimes) <hC> for preglottalized consonants. Otherwise the orthography corresponds to my usage here.

    Yurok has been written a few different ways, including several technical linguistic orthographies and a modern practical orthography. The practical orthography uses <ee, ue, er, e, o, a> for the vowels and doubling to mark length (<eee>, <uue>, <err> for /iː, uː, ɚː/); a following <w> to mark labialization; <sh> and <ch> for /ʃ/ and /tʃ/; <s>, <hl>, and <g> for /ʂ/, /ɬ/, and /ɣ/; and <’> preceding a preglottalized consonant and to represent /ʔ/.

  5. The relationship between Algonquian and Wiyot and Yurok was first proposed by Edward Sapir in 1913 (Sapir 1913) on the basis of fairly weak evidence—there was very little material on Wiyot or Yurok or most Algonquian languages available at the time, nor was there any reconstruction of Proto-Algonquian, and most of Sapir’s equations are now known to be erroneous, though a substantial minority were correct—and was strongly dismissed by the leading Algonquianist of the day, Truman Michelson (Michelson 1914, 1915; Sapir 1915a, 1915b). It remained controversial and mostly unaccepted for decades. The controversy was ended by Mary Haas’s 1958 article which marshaled additional evidence that had become available in the intervening years (Haas 1958), and the relationship completely clinched by Teeter’s (1964a) and Goddard’s (1966, 1975) provision of detailed grammatical connections between the languages.

    Indeed, especially in the absence of substantial numbers of lexical cognates, the most convincing evidence of the relatedness of the Algic languages has been shared, deeply embedded and idiosyncratic grammatical/structural details, what Antoine Meillet referred to as “anomalous forms” (“formes anomales” [Meillet 1925:27], summarized by Campbell 1997:217 as “shared aberrancies”) and Sapir called “submerged features.” A classic example of this sort from Germanic is English’s suppletive bases for the comparative and superlative forms of “good,” /bɛt/ used with comparative /-ər/ and /bɛ/ used with superlative /-st/, perfectly matching German’s /ɡuːt/, comparative /bɛs/+/-ər/, superlative /bɛ/+/-st-/. Goddard has commented that the remarkably specific shared “forms and functions” of the Algic pronominal prefixes “alone would be sufficient to demonstrate a genetic relationship between the languages” (Goddard 1975:250; cf. Alfred Kroeber’s reaction on receiving from Sapir his evidence for the relationship: “The pronouns turn the trick, alone, but the rest looks good” [Kroeber to Sapir, July 30, 1913, in Golla 1984:112], though I think both of these overstate things a bit). The basic most forms of the person prefixes on both nouns and verbs are (1st, 2nd, 3rd, and unspecified, respectively; data slightly simplified) PA *n(e)-, *k(e)-, *w(e)-, *m(e)- : Wiyot ḍ-, kh-, w-, β- : Yurok ˀne-, k’e-, ˀwe-/ʔu-, me-, as in PA *ni·pit- “my tooth,” *ki·pit- “your tooth,” *wi·pit-h/ tooth,” *mi·pit- “a/someone’s tooth” : Wiyot ḍ́əpt “my tooth,” kh́əpt “your tooth,” wəpt́əʔlh/ tooth,” β́əpt “a/someone’s tooth” : Yurok ˀnerpeł “my tooth,” k’erpeł “your tooth,” ˀwerpełh/ tooth,” merpeł “a/someone’s tooth” (Goddard 1975:251). The sound correspondences, except for a few analogical changes in Wiyot, are regular. When preceding an originally vowel-initial non-dependent stem, these prefixes unexpectedly add an epenthetic (*)-t-, as for example PA *ehkwa “louse,” *netehkw[-em-]a “my louse” : Wiyot híkw “louse,” ḍutíkw “my louse.” The Yurok prefixes might appear to have lost this irregularity, but in fact a trace of it remains in the glottalization of the initial consonants, since Yurok changed Proto-Algic **t to /ʔ/ or glottalization in many instances (see, e.g., the “two,” “spruce root,” and “maggot” cognate sets further in the post).

    There are a number of other very specific details involving distribution and morphosyntactic behavior of the prefixes that I don’t have room to present here but which also serve as evidence for the genetic relationship. Note also that when a noun is possessed by a third person in Wiyot, it takes a suffix cognate with the Algonquian obviative singular suffix *-ari (used, among other situations, when animate nouns are possessed by a third person): PA *o-t-ehkw[-em]-arih/ louse (OBV)” : Wiyot hu-t-íkw-əʔlh/ louse.” Other similarities include, e.g., the organization of the numeral and numeral classifier systems, the structure of verb stems and inflection, the use and status of preverbs and the position of person prefixes before preverbs, and several cognate inflectional suffixes and other pronominal elements not noted here.

  6. Noncognate material is enclosed in [ square brackets ]. The reconstructions are my own, though the cognate sets have all been previously proposed by other researchers. For a full list of what I consider attractive or convincing PAc cognate sets, and reconstructions derived from them, see HERE.
  7. Pentland (1983:389) disagrees with this analysis, and reconstructs the diminutive symbolism in this word back to PA. However, while it is admittedly shared by Cree-Innu as well, it is lacking in Ojibwe and Potawatomi (ikwezens and kwézés—with added second diminutive suffixes—rather than xishkwezens and xshkwézés or similar), and other languages are inconclusive. Cree-Innu uses diminutive consonant symbolism far more productively than most other Algonquian languages, and Shawnee and Meskwakian form at the very least an areal if not genetic subgroup, so projecting the diminutive symbolism in this word back to PA, while possible, seems unwarranted.
  8. Some languages maintain the distinction. For instance, some Cree varieties continue the conjunct subjunctive *-e· as , e.g., *nepa·t “if s/he sleeps” → Eastern Swampy Cree nipātē (Ellis 2015:138); and PEA lost most PA final short vowels but retained final long vowels as phonetically short ones in most syntactic environments, a situation continued by languages like Penobscot, e.g., nsəm “my daughter-in-law” ← PEA *nəhθəmPA *neʔθemya vs. nsə̀ma “my absent daughter-in-law” ← PEA *nəhθəmaPA *neʔθemy (Goddard 2007:213-215).
  9. Blackfoot has traditionally been described as a pitch accent or tonal language as well, but recent work by Natalie Weber and others (e.g., Weber 2016a, 2016c; Weber and Shaw 2022) in my view provides good evidence that it is better analyzed as basically having lexical stress which is instantiated primarily via higher pitch.
  10. In general, Algonquian patterns for primary stress assignment seem to be sensitive to the right edge of the word rather than the left. This is little more than a guess, but: given how ubiquitous final vowel or syllable loss is across the family, perhaps primary stress fell on the last stressed vowel before the final vowel of the word?
  11. Furthermore, while Meskwaki in particular is well-documented, Shawnee is one of the most poorly documented and under-analyzed living Algonquian languages; Goddard notes on pg. 44 that “uses of . . . [four of the six Shawnee demonstrative sets] are not completely determined, and a number of forms in sets C and F have not been found in the sources,” and later on the same page refers to a question on whether some members of “set E” are even lexicalized as part of the set as “uncertain in the absence of adequate data on the syntax of their use.” If using only (PEA plus) two non-Eastern languages to arrive at a PA reconstruction, one would hope the reconstructor would use languages with well-understood and fully attested demonstrative systems . . .
  12. Goddard’s explanation of the dramatic changes from his reconstructed PA system to languages like Proto-Cree and thence to individual Cree-Innu varieties is that they involve a combination of “expressive distortion,” of which he gives some examples in various modern Algonquian varieties, to account for things like the vowel lengthening, addition of h+vowel, and replacement of final vowels with *-e·; and irregular phonetic reduction of intial syllables due to their sometimes reduced stress, for which he again provides parallels in other Algonquian varieties where there is more evidence through internal reconstruction or older written records that this is what happened. He argues, reasonably enough, that “[t]he direct evidence for [these processes] within single languages or dialect complexes supports the assumption of the same processes . . . in other languages where there is only comparative evidence” (pp. 80, 82-85).

    Despite Proulx’s (2004b, 2005c) complaints, which are either naïve or in bad faith, both of these are incontrovertibly real phenomena, and reconstructing something like a demonstrative pronoun system, where the words are subject to irregular reductions under lack of sentence stress, as well as to embellishments with various clitics and some instances of “expressive distortion,” not to mention semantic changes, is a monumental task. It’s not Goddard’s fault that language change works this way—my complaint is not with that, but with his decision of how to go about the task of reconstructing the proto-system.

  13. For decades, nearly everyone had assumed that most initial change, or at the very least the process which involved *-ay-infixation, was cognate with a Yurok “intensive” infix /-ɛɣ-/ which was said to also be used in nominalizations and to indicate iterativity, the latter two being more or less the same as the most important reconstructible functions of IC in PA. More recently, however, Garrett (2001) challenged this traditional assumption (Costa [1996:63-64, 70, nn. 76-77] had earlier expressed skepticism and made some of the same points), and it seems that we cannot directly connect the two processes.
  14. TI verbs had three possible theme signs, *-am (“Class 1” = TI(1) verbs), *-aw (“Class 2” = TI(2) verbs), or *-Ø (“Class 3” = TI(3) verbs), the first two of which showed a good deal of allomorphy; unlike in TA verbs, the TI theme signs only vacuously marked the object as inanimate, and each was lexically specified for a given verb. They did not contrast with one another.
  15. The context for this would be when the identities of the verb arguments themselves did not mean that the subject had to be obviative—that is, if the object was anything other than a proximate third person, because in such cases the only reason for the subject to be obviative was because of some triggering proximate which was external to the verb itself. Similarly, the 3OBJ theme sign *-a· could also be augmented with a suffix *-em if a third-person object was unexpected/contrastively obviative, here glossed “MARK.OBV.OBJ.”
  16. Most of the examples are from Bloomfield (1946, 1962), Costa (2001, 2003), Goddard (1974b, 1979a, 1979b, 1981a, 1981b, 1982, 2000, 2001b, 2006, 2007, 2015b), Hewson (1993), and Pentland (1979a, 1999)—making the most use of Goddard (2001b) and Pentland (1979a, 1999), plus Goddard (2000, 2006) for Cheyenne underlying forms—as well as several dictionaries, with the Ojibwe examples usually supplied by myself, as well as often the precise Proto-Algonquian form. (There may be one or two other sources that I unfortunately neglected to keep track of—I don’t think I missed any—but in any case they’d be one of the works listed in the bibliography at the end of the post.) I’ve also supplied a few of the examples from other languages. Any examples constructed by me in a language besides Ojibwe rather than taken verbatim from another source—except for instances that merely involved replacing one person prefix with another—are marked with a double dagger “‡”. In constructing such examples I used Bloomfield (1962) for Menominee inflection, Costa (2003) for Miami-Illinois, Lacombe (1874), Wolfart (1973), and Dahlstrom (1986) for Plains Cree, Ellis (2015) for Moose and Eastern Swampy Cree, Goddard (1979b) for Munsee, and Goddard (1994a, 2004) for Meskwaki, as well as my own background knowledge and some reasonable extrapolations.

    Note that in at least a few cases, the cited PA form likely did not exist as such, and is really the result of projecting parallel structures which arose later in several daughters (or which now exist with the specific morphemes in question, but didn’t previously) back to a pseudo-PA form. All of the phonological and morphophonemic processes such pseudo-PA and daughter forms are being used to illustrate are valid, however.

    Additionally, in some cases one or more daughter forms used in an example don’t reflect a complete, regular descent from the cited PA form, due to various analogies, the need to use additional particles to make a grammatical construction, etc.; these are sometimes simply ignored, while in other cases I enclose the non-cognate material in [square brackets].

  17. Pentland (1999:249, 251) offers some weak evidence that *-n(ay) had some additional realizations as well, specifically: (1) *-(e)na- before the 2pl/3pl pluralizer *-wa·w (by Rule 8 below), which would be consistent with *-wa·w in most other circumstances not inserting a preceding epenthetic vowel, but is only potentially reflected in Menominee, and even there a good alternative explanation is available (Pentland 1999:249, n. 38, citing p.c. from Goddard; Goddard 2007:244, n. 50); and (2) instead of losing *-ay before all vowels, the *-ay combined with a following *-a-initial suffix and contracted to *-a·-, potentially reflected in Delawaran, Massachusett, and the Kitigan Zibi Nipissing dialect of Ojibwe. Again, however, the distribution of this reflex is not widespread, has at best been disturbed by various reshapings, and has other explanations available (Goddard 2007:261-263), although I confess I don’t find Goddard’s explanation entirely satisfying either.
  18. This rule and consequently several others are here given slightly differently from how they have traditionally been presented. Bloomfield (1946:93, 101, 102; only implicitly for some of these statements) and later Algonquianists saw *-y-insertion as occurring specifically: (1) between long vowels, (2) in AI conjunct forms between a verb stem ending in a vowel and a central agreement suffix beginning with a vowel, and (3) in conjunct forms between the 1OBJ theme sign *-i and a following vowel-initial suffix. In other cases, a vowel next to another vowel was lost. In such cases, the vowel which dropped was always the short vowel in any sequence containing a short and long vowel in either order, and otherwise it was, as Pentland (1979a:406) put it, “the one farther from the centre of the word.” But this last rule has been leveled out in some daughter languages, where it is always the second of two consecutive short vowels that drops: e.g., the example from Rule 2a above of *ke-2” + *-atay- “belly, stomach” → *katayi “your belly, stomach,” reflected in Plains Cree katay and Munsee kătay, but where Passamaquoddy has leveled the rule out in kə̀t (form from LeSourd 1988:69; PMD).

    It would be simpler, however, to just combine all the *-y-insertion rules into one, and dispense with most of the vowel deletion rules—most of the short vowels to be “deleted” being the epenthetic ones, which would simply never have been inserted to begin with. This is Proulx’s (2003a:222, n. 4; 2005b:197 and n. 5) argument. This gives the correct outputs in most cases, but in at least some instances there are processes in the modern languages which don’t adhere to this rule and seem to be reconstructible to PA. At least descriptively, I’ve chosen to basically follow his simpler but not-quite-accurate option, in which case we would have to define Rule 5 such that it has certain given exceptions (as with Rules 1a and 14a versus 1b and 14b).

    (The real reason I’ve chosen to do this is that I really do not have the time or patience to rewrite all these stupid rules any other way, sorry.)

  19. “Stone, rock” is one of several words or word classes which different Algonquian languages, and even dialects within the same language, assign to different genders, so it’s not entirely clear if it was animate or inanimate or dialectally both in PA. Bloomfield was actually inconsistent in his reconstructions: he gave the plural form as inanimate <*aqsenyali> = (properly) *aʔsenye·ri (1946:86, #8) but the singular as animate <*aqsenya> = *aʔsenya (1946:93, #87). Perhaps, as in some daughter languages like Meskwaki (Goddard 2002b:214, 216), “stone, rock” was generally inanimate, but was animate when it referred to special, unusual, important, and/or more specific kinds of stones/rocks, and this latter usage was extended to treating the noun as animate in all instances in many daughter languages.
  20. Rhodes (2021:310ff) follows Goddard’s (1967) original reconstruction of the distribution of the allomorphs of *|am|, rather than his more recent one (Goddard 2007; Goddard’s more recent reconstruction is in this respect shared with Pentland 1999) which I adhere to here. Pentland’s and Goddard’s later reconstructions have better support and result in more logical overall protosystems.
  21. In fact, other than the metathesis of *|aw-k|, Rule 16b seems to basically just be the operation of Rule 9 (see the discussion of *|awe|-contraction below), including in most cases *|aw-w| → *|aw-e-w| → *a·w, but it’s convenient to treat it as a separate process here.
  22. Goddard (2001b:208) also reconstructs it as contracting to *a· before *-etwiRECIP,” leading to *-a·twi (*|-aw-etwi|), but the variant formation *-āwatī found in PEA (rather than x-ātī) is unmotivated and looks archaic. Proulx’s (2003b:411-412) suggestion, that PA had *-a·watwi for this combination, and the *-a- was replaced by *-e- analogically and then the *a·we sequence was contracted to *a· in most of the non-Eastern daughter languages, seems plausible.

4 thoughts on “Proto-Algonquian

  1. This blog and especially this “freaking monograph” is wonderful — thank you for all the work that went into this synthesis, which is a great contribution to Algonquian linguistics. Guides to the state of the art like this are few and far between outside of Indo-European; in fact, going through the literature bit by bit myself to familiarise myself with the field, I was musing to myself over the past week about writing just such a synthesis myself. You should really consider eventually expanding this and your posts to come into a published text, either on your own or eventually collaborating with another Algonquianist or Algonquianists.

    Just one obvious question: I and others (cf. a post yesterday to the Algonquiana mailing list by Monica Macaulay) have been wondering who the author of this blog is? This both out of curiosity and because this blogpost in itself is so obviously citeable.

    Liked by 1 person

  2. In response to Christopher Ray Miller’s comment above, I have to agree that this blog is very interesting and those of us on the Algonquiana mailing list would be very interested to know the identity of the author responsible for this work, and I would love for his or her work with Algonquian linguistics to continue. Gichi-miigwech!


  3. Boozhoo ndinawemaagan! If you’d ever like to get into a discussion about Algonquian speakers and archaeological cultures, I’m another person who thinks about that all the time, but I’m more on the Archaeology side of things! ;)


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s