On the structure of (personal) pronouns in Inuktut

Abstract This article examines the internal structure of personal pronouns in Inuktut. Building on previous work on syntactic variation and universals in personal pronoun systems (Harley and Ritter 2002, Ghomeshi and Massam 2020) and drawing on arguments that these and demonstrative pronouns are structurally complex (Déchaine and Wiltschko 2002, Leu 2015), it is argued that Inuktut pronouns are morphosyntactically complex, with exponents realizing multiple functional heads as well as an overt root. In particular, it is claimed that (local) person is represented twice, both as a root and in a higher functional projection. Previous arguments that Inuktut pronouns should instead be analyzed as D heads based on Adnominal Pronoun Constructions are also addressed (Yuan 2018, 2021).


INTRODUCTION
Since Postal (1969) argued that English pronouns are essentially definite articles, pronouns cross-linguistically have often been analyzed as (or assumed to be) determiners. 1 With the adoption of the DP hypothesis (Abney 1987), this came to mean that pronouns were treated as intransitive determiners, lacking an NP complement. For instance, Ritter (1995) argues that first and second person pronouns in Hebrew are D heads. Our conception of pronouns has been further refined by Déchaine and Wiltschko (2002: 410), who argue that the size of pronouns varies cross-linguistically (and even language-internally), with pronouns being instantiated as pro-DPs, pro-f Ps, and pro-NPs, as shown in the structures in (1).
Déchaine and Wiltschko's evidence for different-sized pronouns includes the syntactic distribution of pronouns in a given language (e.g., whether they appear in argument or predicate positions), their binding possibilities (as fPs are argued to be able to act as bound variables, while DPs are instead argued to act as R-expressions), and their morphological properties, such as their ability to combine with determiners (see Moskal 2015 andSmith et al. 2019 for other decompositional approaches to pronouns).
Inuktut 2 pronouns not only corroborate this type of decompositional approach to the structure of pronouns, they appear to extend it; realizing both the exponents of an extended functional projection and a (nominal) root which is often considered to be null or elided in languages with pro-DP and pro-fP pronouns (Ritter 1995;Déchaine and Wiltschko 2002;Elbourne 2001Elbourne , 2005Merchant 2014). In particular, I argue that since pronouns conform to the word-initial root requirement of words of the language, they can undergo noun incorporation as expected of nominals (containing a 1 The following abbreviations are used: ABS: absolutive case; ABL: ablative case; ALL: allative case; APC: Adnominal Pronoun Construction; ASP: aspect; ASSOC: associative; COP: copula; CTMP: contemporative clause type; DECL: declarative/participial clause type; DEM: demonstrative; DIST 2 Inuktut is the term proposed by Inuit Tapiriit Kanatami for the Inuit language in Canada. The more commonly known term 'Inuktitut' refers to dialects in the eastern Canadian Arctic. root), and they exhibit categorial flexibility insofar as wh-pronouns can also serve as predicates, which is consistent with pronouns (of all types) containing overt roots.
In addition, pronouns in Inuktut appear to constitute a counter-example to the generalization made by Ghomeshi and Massam (2020) that pronominal number is dependent on person. 3 Instead, number-marking on Inuktut pronouns largely patterns as on common nouns. I suggest in what follows that this is tied to the fact that Inuit pronouns contain roots and are thus noun-like.
Finally, I consider arguments by Yuan (2018Yuan ( , 2021 based on Adnominal Pronoun Constructions (APCs) that Inuktut pronouns are instead D heads, and provide alternative analyses of these forms that suggest they are not in fact APCs. The main empirical and theoretical claims of the present article are as follows: (2) Inuktut pronouns: a. are full DPs, with an articulated structure realizing multiple functional projections, b. realize person features twice: on both a dedicated root and a higher π (or f) projection (which interacts with number and case, suggesting it is part of the same functional sequence), c. contain an overt root, instead of an elided expression, as typically found in languages with pro-DPs and pro-fPs, d. are noun-like insofar as their number marking patterns with common nouns. (Ghomeshi and Massam 2020) In the next section, I begin by presenting some background on Inuktut pronouns. Then I present arguments that these pronouns instantiate multiple syntactic projections and include a root, drawing on evidence regarding Inuit word structure, the ability of pronouns to undergo noun incorporation, and additional aspects of the behaviour of demonstrative and wh-pronouns.

BACKGROUND ON INUKTUT PRONOUNS
Inuktut has distinct (strong) personal pronouns only for first and second persons, as illustrated in Table 1, with data from Inuinnaqtun (Western Canadian Inuktut) (Lowe 1985a: 94-96); various demonstratives (such as una 'this one') are used for third person referents. 4 As with nouns in the language, three numbers are distinguished: singular, dual, and plural. 5 Pronouns also inflect for case, as discussed in section 3.1. 6 While Inuktut is often characterized as a pro-drop language, it is primarily ergative and absolutive personal pronouns that are omitted, arguably because these same arguments have their phi-features coindexed on the verb (see Yuan 2018, and references cited therein). Personal pronouns routinely surface in oblique cases, as in (3a); in constructions without a verb, as in (3b); and for focus, as in (3c) Applying the diagnostics of Déchaine and Wiltschko (2002) to Inuktut personal pronouns, they appear to behave as pro-DPs. In particular, they have the distribution   (Lowe 1985a: 96, glosses added) Like the other pronouns examined herein, it appears to have an articulated structure, containing the reflexive morpheme -mi-, found on third-person possessed nouns bound by their subjects, as well as possessor agreement. 5 While this three-way number contrast holds for most dialects in Canada, the South Baffin dialect has lost the contrast between dual and plural on both verbs and nouns, including pronouns, as have Greenlandic dialects. 6 The second person plural pronoun iliffi, presented here in the Inuinnaqtun orthography, is given in subsequent examples in its underlying form, /iliphi/. 7 Glosses have been adjusted for consistency with the rest of the data in this article.

197
COMPTON of arguments and can only act as predicates when accompanied by a copula, which in turn hosts tense, clause-type marking, and agreement. 8 As noted above, demonstratives can be used for third persons. The language has a rich system of demonstrative pronouns, whose inventory varies by dialect. Western dialects have a particularly extensive set of contrasts, as exemplified with data from Sallirmiutun (formerly called Siglitun) in Table 2. While absolutive singular is null on nouns, it is marked by -na on demonstratives, appearing across all (and only) absolutive singular forms (as has also been reconstructed for Proto-Inuit-Yupik forms by Fortescue et al. 2010: 499-526), as well as some wh-items. 9  The primary contrast throughout the demonstrative system involves whether the referent is visible (abbreviated as VSBL). This is followed by a distinction between referents that are more physically restricted (RSTR) in their scope and can be pointed to, as opposed to referents whose shape is more extended or whose boundaries are ill-defined (Lowe 1985b: 272). 10 Next, some demonstratives are specified for proximity to the speaker (first person), the listener (second person), or distance from both of these (marked by -), whereas others are underspecified with regards to distance from speech act participants. Finally, some demonstratives specify a relative spatial position above, below, inside, outside, or across. Such meanings are at times idiosyncratic, indicating objects located "upstairs", "on the roof", or "by the door; that one in the cold porch", etc. (Lowe 1985b: 276-278). Additional forms (not shown) mark dual and plural number and the seven remaining morphological cases beyond absolutive.
While demonstratives can be used pronominally, as in (4a), they can also cooccur with a nominal restrictor, as in (4b). Although this might lead us to conclude that they are demonstrative determiners, two factors suggest otherwise. First, they can act as possessors-a property of nouns in the language-bearing a different case from the possessed noun, as shown in (5a). Second, the can be discontinuous from the nouns they modify, as illustrated in (5b), where the demonstrative is separated from a noun by the interjection haa 'look at'.  Compton and Pittman (2010) and Compton (2012), I assume demonstratives with a nominal restrictor bearing the same case to be two DPs in apposition.
Finally, the language has a set of wh-pronouns, such as suna 'what', kina 'who', and qanuq 'how', which also inflect for case and number and whose behaviour we return to in section 3.2.

EVIDENCE FOR INTERNAL STRUCTURE
This section presents evidence that Inuktut pronouns do not merely instantiate a single functional head, but instead instantiate multiple syntactic projections, including an overt root. While the focus is on personal pronouns, additional arguments are drawn from the behaviour of demonstrative pronouns and wh-pronouns.

Inuktut pronouns are multi-morphemic DPs
A first observation about Inuktut personal pronouns is that they contain two main parts: an initial stem indicating person (argued to be a root in the next subsection), followed by a series of functional morphemes marking person (a second time), number, and case. These two domains can be observed in Tables 3 and 4, where the various forms of the first and second person pronouns in Inuinnaqtun are presented. 11 Note that while ergative and absolutive forms are syncretic here, the two cases are distinguished elsewhere in the language (e.g., on singular unpossessed nouns, on possessed nouns, on demonstratives, and on wh-words).
The only morpheme common to all first-person forms is the stem uva. Similarly, only the stem il(i) appears in all second-person forms. Furthermore, in some dual oblique-case forms, these morphemes are the only element distinguishing first and second person. As such, we can conclude they contribute FIRST PERSON and SECOND  (Harley and Ritter 2002). That a root might encode person is not entirely surprising, given that Déchaine and Wiltschko (2002) argue that Japanese kare 'he' is a pro-NP consisting of a noun.
The remaining morphemes co-vary with person, number, and case (with some instances of syncretism between dual and plural) and, in fact, appear in other contexts. The absolutive and ergative endings occur as agreement on verbs, as illustrated in (6) (Lowe 1985a: 108). 12 In particular, k and t are default realizations of dual and plural number in the language, appearing on verbs in the third person, as in (7), but also on bare unpossessed absolutive nouns (e.g., arnaq∼arna-k∼arna-t 'woman, two women, more than two women') (Kudlak amd Compton 2018: xxiii). 13 (6) a. niri-yu-nga (7) a. niri-yuq eat-DECL-1(SG) eat-DECL.3SG  (2017) argues instead that -gu(t) in first person dual and plural pronouns is the associate marker -kku(t) found elsewhere in the language and discussed in section 4. Herein I take its absence from second person forms to support a different analysis. Fortescue et al. (2010: 418) note in their entry for the reconstructed proto-form *uvakut 'we' that "the kut seems to be the same as the 1pl verbal subject marker of the same shape" (see, e.g., their p. 489). Corbett and Mithun (1996: 16) also compare the equivalent forms in Yup'ik, but ultimately suggest the similarity may be a coincidence. 13 Note that, like many suffixes in Inuktut, dual -k and -t delete a word-final consonant, such as q. The plural form -t in Inuinnaqtun co-exists in variation with the phonologically conservative form -n.

COMPTON
The oblique form endings of pronouns also occur in other contexts, matching almost exactly the endings on the corresponding possessed nouns. 14 For instance, the first-person endings mark a first-person possessor on nouns, as in Table 5.
On some singular oblique possessed noun forms (as well as the corresponding pronouns), a velar consonant -ng-marks second person, as contrasted with a labial consonant -m/p-marking first-person forms, as shown in (8). 15 However, in other forms, this contrast is neutralized to the labial consonant, yielding forms ambiguous between first and second person, as in (9). For concreteness, I assume this syncretism to be due to feature impoverishment (Bonet 1991, Halle 1997, removing the feature HEARER and leaving the feature PARTICIPANT. 16 (8) a. tupi-m-ni b. tupi-ng-ni c. tupi-a-ni tent-1-LOC tent-2-LOC tent-3-LOC 'in my tent' ' in your tent' 'in his/her tent' This decompositional approach yields to the following descriptive schema for Inuit personal pronouns. An initial stem marking person, followed by an independent morpheme marking person (as further evidenced by possessive forms), a morpheme marking non-singular number, and a morpheme marking oblique cases.   14 The vialis and similaris case dual and plural endings, distinguished by a geminate consonant when part of the first person pronoun, are syncretic as possessor marking on nouns. 15 Though not decomposed in this way, traces of this marking can also be seen in reconstructed case forms in Fortescue et al. (2010: 487-488), who attribute the origin of m to ergative (their relative) case. 16 That -m/p-is the exponent of the feature PARTICIPANT is further evidenced by the fact that while it appears with both first and second person forms, it never arises as part of third person possessor forms (the reflexive -mi-being unrelated). Similarly, -ng-marking second person is also absent from third person endings, although a dual number marker -k may surface as ng due to assimilation. 202 CJL/RCL 67(3), 2022 (10) Template for Inuktut personal pronouns ffiffiffi π p À π À ð#Þ À ðD=KÞ Abstracting away from cases where feature impoverishment removes certain features, and setting aside pronoun roots for the moment, I propose the following structures (adapted from Ghomeshi and Massam 2020) and Vocabulary Insertion rules for both personal pronouns and their corresponding possessor marking on nouns. 17 For convenience, I use both SPEAKER and HEARER features here in lieu of bivalent features (see, e.g., Nevins 2011, for  The proposed structure for pronouns is similar to that of Moskal (2015: 365, ex. 6), who also posits that # occurs higher than person (which is instead realized on a lower D head), but without pronouns containing a root. For expository purposes, Smith et al. (2019Smith et al. ( : 1056Smith et al. ( -1057 treat the pronominal bases in Moskal (2015) as roots instead of D heads, but note that they "do not intend to take a stand on whether pronouns have roots" (Smith et al. 2019: 1036, note 10). 18 Little n is omitted from alongside the nominal for expository purposes. I abstract away from the possible distinction between D and K. While the number marking shown thus far on possessed nouns tracks the possessor, and thus might well co-occur with person features on the Poss head, the fact that these number contrasts are neutralized/bled by non-singular number of the possessum noun suggests they compete for insertion on a single head. For instance, iglu-vuk 'our two houses' is compatible with either a dual or plural possessor. 19 While there is overlap between the Vocabulary Items marking f-features in pronouns and the possessor agreement forms on nouns, they are not always identical. In some cases this may be due to context-sensitive allomorphy or to the fact some possessor agreement forms on nouns are portmanteau, bearing two bundles of phi-features indexing both the number of the noun and the f-features of the possessor (see, e.g., Oxford 2019 on portmanteau agreement). 20 Given that -hi-alone marks second person plural agreement on verbs, it may instead constitute a span or instance of fusion marking both person and number. Crucially, personal pronouns are multimorphemic and structurally more complex than might be expected of simple determiners.
In the next subsection I argue that the stems of these pronouns are roots.

Inuktut pronoun stems are roots
A number of disparate phenomena point to Inuktut pronoun stems being roots as opposed to purely functional projections.
Beginning with the most general evidence, it has been observed in both descriptive and theoretical work on the language that words in Inuktut follow the schema in (14) For instance, nominals and verbal complexes in Inuktut must normally begin with a root (Johns 2007). This includes instances of noun incorporation, where a closed class of suffixal verbs obligatory incorporate a noun, as in (15) where the NI verb -tu-'consume' must incorporate its internal argument. Similarly, a closed class of suffixal restructuring and modal verbs can also incorporate a verb (Woodbury and Sadock 1986, Johns 1999, Pittman 2009), as in (16) where the restructuring verb -rqu-'order' combines with the verb oqalo-'speak'. (West Greenlandic; Grimshaw andMester 1985, Woodbury andSadock 1986: 238, adapted) 21 An alternative analysis could posit that the ng portion of these morphemes was part of the subsequent oblique case markers. I set aside this possibility since this segment is absent from both third person possessive forms and unpossessed nouns in oblique cases. The possibility of ng in dual forms being an assimilated k is a potentially confounding factor, but cannot account for plural forms. 22 Two possible exceptions to this generalization include instances of ta-(the only prefix in the language), which arises only in the demonstrative system, and a phenomenon of stem ellipsis in Nunavik Inuktitut akin to VP-ellipsis where a verbal stem can be omitted when contextually/discourse salient (Swift and Allen 2002).

204
Given the fact that modals, restructuring verbs, and incorporating verbs in the language are argued to be functional in nature (Johns 2007, Cook andJohns 2009), this means that these elements cannot be pronounced in isolation, as they lack a root. Instead, a dummy root pi 'do, get, thing' is inserted in such contexts; for instance, appearing in dictionary entries to render such forms pronounceable: (17)  That pronouns need no such dummy root to satisfy the requirements for wordhood in the language is explained if they are not solely comprised of functional material, but instead already begin with a root. Another phenomenon involving roots (and larger projections containing them) is incorporation. Closed classes of verbs obligatorily incorporate either nominal or verbal complements-both built up from a root (see, e.g., Johns 2007, Pittman 2009, and references cited therein). For instance, the copula in Inuktut must incorporate its nominal complement. However, this requirement can be satisfied with a pronoun, as in (18)  The ability of pronouns to undergo incorporation also extends to demonstrative pronouns, as in the following example. '… was one of [those we had]' (Baffin Inuktitut;Farley 2012Farley : 2002 Thus, like common nouns in the language, pronouns can also undergo incorporation into verbs that obligatorily incorporate a nominal complement, further suggesting that they too contain roots. Pronouns may also undergo modification by suffixal elements that typically modify nouns, such as -ruluk 'bad' in the following example: Yet another root-like behaviour among pronouns in the language is found with wh-pronouns. In addition to their use as nominal wh-expressions, many of these stems can also serve as main predicates in a clause, taking the tense, clause-type, and agreement morphology normally found with verbal roots, as illustrated in (22) and (23). 24 24 For concreteness, I assume these are acategorial roots which combine with little n when used as nominal wh-expressions and little v when used as verbal predicates. As noted by Sherkina-Lieber (2004), wh-expressions in Inuktitut can also undergo incorporation, further demonstrating that their behaviour patterns with nouns: 'What do you want to have?' (Sherkina-Lieber 2004: 126, adapted) While it was assumed earlier that -na marked absolutive singular on nouns, as its presence covaries with case and number in Western dialects (Lowe 1985a,b,c), its appearance on incorporated suna 'what' here might instead suggest it to be a form of stem allomorphy. It could also be the realization of a little n categorial head, as it only appears in nominal and not verbal contexts. It is also possible for some wh-expresions to occur with a nominal restrictor, in which case they nevertheless bear their own case and number marking. Compton and Pittman (2010) and Compton (2012) argue that such constructions in fact involve two DPs in apposition. While in such cases it is not possible for the wh-expression to incorporate, such an expression may nonetheless modify an incorporated noun. In sum, the ability of wh-expressions to appear as both nominal and verbal is consistent with their being acategorial roots whose category is determined by their syntactic environment. 25 As such, they too display root-like behaviour, suggesting a more general property of Inuktut is that pronounsof all typescontain roots. 26 A final observation about wh-expressions in Inuktut is that we might expect a Dlinked wh-expression like naliak 'which' to be a determiner, given that D-linked whitems are presuppositional and such information is often encoded by determiners (Pesetsky 1987). 27 Crucially, unlike other wh-words which simply inflect for case and number, this wh-item also bears possessive morphology, as noted by Lowe (1985a,b,c): (24) nali-anni it-pa? which-3PL.POSS:LOC.SG be.located-INTERR.3SG? 'In which one (of many) is it?' (lit. 'in their which-one is it located?') (Sallirmiutun ;Lowe 1985b: 110) That naliak 'which' bears the possessive marking normally found on nouns with third-person possessors is consistent with it too being nominal, and not of category D. Moreover, while naliak can combine with a nominal restrictor, it once again 25 An additional function of wh-items is to serve as a base to create indefinite pronouns.
(i) nak(k)ikkiaq nakkit-kkiaq where.ABL-uncertain 'I don't know from where; from somewhere' (Spalding andKusugaq 1998, Pirurvik Centre 2020) 26 That pronouns behave like nouns coincides with the descriptive literature that claims words are either nouns or verbs, with a small class of particles (Fortescue 1980). 27 Thank you to an anonymous reviewer who pointed this out. Such examples suggest the existence of two DPs, which is again consistent with naliak containing a nominal root.

Pronominal number patterns like grammatical number on nouns
Part of deciphering the structure of Inuktut pronouns involves isolating the position of number. In a study of the morphosyntactic position of number in pronominals based on data from Persian and Niuean, Ghomeshi and Massam (2020: 604) make the following claims (amongst others regarding, e.g., proper nouns): (26) a. Grammatical number in common noun phrases is part of the nominal spine […], similarly to TP within CP, allowing for individuation, hence for referentiality.
b. Pronouns do not include a number position in their spine […]. Number in pronouns never plays a role in individuation; rather, person does.
c. Pronominal number is always bundled with π (or it is attached to another position within the pronoun) and it is always subordinate to π, which is spinal.
Ghomeshi and Massam argue that number is both syntactically and semantically distinct in pronouns and nouns, proposing that pronominal number is expected to be bundled with person. For instance, they argue that (bundled) pronominal number may be used for social deixis in Persian (see also the tu-vous distinction found in French), with a higher (grammatical) number projection being added at times to disambiguate "ambiguities created through the honorific system" (Ghomeshi and Massam 2020: 601).
Interestingly, Inuktut seems to be an exception to generalizations (26b) and (26c), insofar as at least the first-person pronoun exhibits an identical pattern of number marking as is found on unpossessed common nouns, demonstrative pronouns, and wh-pronouns, as in Table 6 (showing absolutive forms)all of which arguably lack (marked) person features (Ghomeshi and Massam 2020: 601). 29 28 A reviewer notes that the form apqut-it is in principle ambiguous between ABS.PL and ERG. PL and suggests naliatigun and apqutit could be in apposition. However, given that the inflectional ending -atigun indicates a singular referent with a plural possessor, it seems unlikely that there would be a number mismatch between appositives. Unfortunately this is difficult to test, as we would not predict the unambiguous ergative singular form -up to be able to occur on nominal restrictors of naliak, for the same reason that *which of the road is not possible in English. Lowe only gives examples with dual an plural restrictors. 29 Pronouns, demonstratives, wh-words, and also common nouns display varying degrees of allomorphy, as illustrated by the stem allomorphy in the demonstratives shown here.
As with common nouns in Inuktut, grammatical number occupies a position in the functional spine of the DP between the root and lower exponents of person, on one hand, and higher case-marking (on D or K), on the other. Even in more complex forms that are affected by allomorphy, such as the oblique pronouns in Tables 3 and 4, the set of distinct exponents (i.e., for person and number) and their relative order, as in (27) repeated from (10), suggest that person and number occupy distinct syntactic positions, with number realized on a higher projection, as might naturally be expected given the Mirror Principle (Baker 1985).
(27) Template for Inuktut personal pronouns ffiffiffi π p À π À ð#Þ À ðD=KÞ This relative order also coincides with the fact that the exponence of number is conditioned both by person morphology and by case, as expected if it occupies an intermediate position between the two. For instance, the forms ti({ng/k})/hi({ng/k}) are the exponents of GROUP number after -p-(marking PARTICIPANT), but show further conditioning according to case (see, e.g., the plural forms of the second-person pronoun in Table 4).
I propose here that the exceptionality of Inuktut pronouns with respect to the generalizations proposed by Ghomeshi and Massam (2020) is due to the fact that they exhibit a structure more similar to common nouns: they contain a root and the sequence of functional projections typically found with nouns, including an independent # projection.

COUNTER-EVIDENCE FROM ADNOMINAL PRONOUN CONSTRUCTIONS
As part of a much larger analysis of (split) ergativity and f-marking in Inuktut, Yuan (2018: 152) proposes that "independent pronouns in Inuit are bare D 0 s, not phrasal DPs" (see also Yuan 2021). Evidence for this claim is drawn from Adnominal Pronoun Constructions (APCs), following work by Postal (1969) that argues that (plural) pronouns in English may combine with nouns and in such constructions are in complementary distribution with articles. Yuan (2018)   In addition, Yuan (2018) proposes that the following examples from Inuktitut also constitute APCs, analyzing the suffixes as pronominal clitics (and observing parallel forms of agreement on verbs).
(30) a. ilisaiji-tigut teacher-1PL 'We teachers' (North Baffin, Arctic Bay) b. ilinniaqtu-hi student-2PL 'you (pl.) students' (Kivalliq, Arviat;Yuan 2018: 153) Beginning with the Kalaallisut example in (29a), there are several factors that call into question whether this is indeed an APC involving a bare D 0 combining with an NP. Firstly, the word kalaalliit in (29a) includes the suffix -(i)t, marking plural number and case (albeit ambiguous between absolutive and ergative), and thus would surface identically as a full DP without the pronoun. Crucially, full DPs can occur in apposition with each other, as noted by Fortescue (1984) and illustrated in (31). 31 (31) a. arnaq kalaaliq woman.ABS.SG greenlander.ABS.SG '(the/a) woman who is Greenlandic' b. kalaaliq arnaq greenlander.ABS.SG woman.ABS.SG '(the/a) Greenlander who is a woman' (Kalaallisut; Fortescue 1984: 51) 30 The abbreviations used in glossing Yuan's examples, particularly for SG and PL, have been modified in some cases for reasons of consistency, but the content has not been changed, unless otherwise noted. 31 More generally, relativization in the language employs nominalized clauses in apposition with a head noun. 210 CJL/RCL 67(3), 2022 Given the extensive use of DP-DP appositives in the language, the proposed APC in (29a) might well involve two full DPs in apposition. 32 Turning to the Inuktitut examples in (30), as Yuan (2018) observes, these same suffixes appear as f-marking on verbs. Setting aside their status as clitics or agreement (although see Compton 2014, Yuan 2015, Compton 2017, Johns and Kučerová 2017, and Yuan 2018, 2021 for discussion), we can first note that they are formally distinct from the independent pronouns that are the focus of this article (although they appear to instantiate a subpart of these pronouns). However, in addition to this, there is reason to believe that they too are not APCs. Crucially, the ending -tigut in (30a) is in fact the expected vialis case first-person plural possessive form (cognate with the Inuinnaqtun form -ptigun in Table 5, but also identified as -(t)tigut for Inuktitut by Dorais 1988: 36, who calls it translative case). 33 Moreover, while the expected form of 'teacher' is iliniaq-ti, with an agentive nominalizing suffix, the form in (30b) appears to instead involve the intransitive declarative/participial ending -tu(q). As such, I propose the alternative glosses and translations in (32), where (32a) is a possessive construction where the vialis case adds the meaning 'among', 34 and (32b) is either a