1 Introduction

Suppletion refers to the phenomenon where a single morphological item is associated with two phonologically unrelated forms, the choice of form depending on the morphosyntactic context. Canonical items from English include good-better-best, bad-worse-worst and go-went. Suppletion has been shown in recent work to be quite regular in the sense that although we may not be able to predict which exact lexical item in a given language will be suppletive, we can predict in which ways suppletion is constrained. For instance, which elements can serve as a context for suppletion have been shown to be delimited in terms of locality restrictions (see Adger et al. 2003; Embick 2010; Bobaljik 2012; Moskal 2015a; b; Merchant 2015; Moskal & Smith 2016, i.a.).

In addition, we have a good idea of what kind of suppletive patterns we can expect. For instance, in adjectival suppletion, Bobaljik (2012) shows that in positive-comparative-superlative triples, the only patterns found are the following: AAA (no suppletion), such as long-longer-longest; ABB (where the comparative and superlative supplete together), such as bad-worse-worst; and ABC (where the comparative and superlative are both suppletive relative to the positive, but also with respect to each other), such as in Latin bonus-melior-optimus ‘good-better-best’. AAB patterns (hypothetical ‘good-gooder-best’), where only the superlative is suppletive, are not found, nor are ABA patterns (hypothetical ‘good-better-goodest’), where the comparative suppletes on its own. Looking beyond adjectival suppletion triples, AAB patterns are not universally unattested, however. Taking triples to represent markedness relations of increasing complexity (to be made more precise below), AAB patterns have been found in other suppletive contexts, such as case and number (see Smith et al. 2018) as well as clusivity (see Moskal 2014). In contrast, a remarkably strong result is that ABA patterns are almost universally unattested: in addition to the lack of ABA patterns in adjectival suppletion in Bobaljik (2012), they also do not seem to be attested in case or number (Smith et al. 2018) or in syncretism in pronouns (Vanden Wyngaerd 2018). In this paper, I add another set of data that supports the observation that ABA patterns do not arise, namely in clusivity-driven suppletion. This observation then leads to a morpho-syntactic analysis of clusivity, which crucially relies on privative features.

The structure of the paper is as follows. In section 2, I first provide a brief overview of clusivity and introduce the main observation of the paper: the lack of ABA suppletion patterns in clusivity. Next, I briefly summarise how suppletion serves as a tool for identifying structural representations (Bobaljik 2012; Smith et al. 2018). In section 4, I analyse the main clusivity suppletion patterns, showing how ABA is underivable in the proposed system and, as such, predicted to be unattested; section 5 discusses apparent counter-examples to this claim. Section 6 deals with the issue of markedness and its role in suppletion. The term markedness (Trubetzkoy 1939; Jakobson 1941) has acquired several notions and a loaded meaning.¹ It may be helpful to note that in this paper I address the following three senses explicitly: (i) structural markedness: one element is more marked than another if its morpho-syntactic structure is more complex; (ii) featural markedness: in a feature that has two values, one of the values is more marked than the other; and (iii) overt coding (“morphological markedness”): when only one exponent of a particular feature (value) is overtly (phonologically) expressed, this feature (value) is more marked.² In section 7, I discuss the consequences of the proposal here, and compare it in particular to assumptions about the representation of person in Harbour (2016). Section 8 concludes the paper.

2 Clusivity

As visualised in (1), the inclusive/exclusive distinction captures the difference whether the addressee (or addressees, represented by 2) are included or excluded from the set of referents which also contains the author, 1.³ Note that 3 represents those who are neither author nor addressee (i.e. third person referents).

When the inclusive is used, the addressee is crucially included, while the exclusive indicates that the addressee is excluded. This is a distinction that is frequently seen across languages (Cysouw 2003; Filimonova 2005; Siewierska & Bakker 2005). While the inclusive/exclusive distinction can be found in verb agreement affixes and possessive affixes as well, I here limit myself to inclusive/exclusive plural independent pronouns and focus on a binary number distinction between singular vs. non-singular.⁴ The pronominal database consulted for this project is the Free Personal Pronoun System database⁵ (Smith 2011), which encompasses 455 languages in total. In the Appendix, I list the subset of languages from the database that make a clusivity distinction, along with the judgements that I assign to them.

As mentioned above, morphological marking of the distinction between inclusive and exclusive (first) person (plural) is relatively frequent cross-linguistically (Cysouw 2013) and either the inclusive form or the exclusive form can be overtly coded (Harbour 2016). Indeed, in (2), we see that in Itzaj Maya (Hofling 2000) the inclusive contains the sequence -e’ex in addition to the exclusive form. In contrast, in Limbu (van Driem 1987) it is the exclusive that is expressed by an additional sequence -ge on top of the inclusive form (see (3)).⁶

However, as noted by others, the first person inclusive has been proposed to be a marked category (Noyer 1992; Cysouw 2003; Siewierska 2004; LaPolla 2005, a.o.). For instance, in accordance with the diagnostics in Greenberg (1966) for marked categories involving at least as many morphemes as the corresponding unmarked categories, while exclusive marking is attested, inclusive marking seems to be more common (cf. Harley & Ritter 2002). It is important to stress, though, that whilst there may be a trend towards overt coding of the inclusive as opposed to the inclusive, both types of clusivity can be overtly coded (see Harbour 2016, i.a.) (see also section 7 on the role of overt coding).⁷

Another asymmetry between the inclusive and exclusive is noted by Cysouw (2005; 2013). There are a few languages that have a special pronoun for the inclusive, but the pronoun used to express the exclusive is identical to the first person singular pronoun (see also Sokolovskaya 1980); for instance, in Canela-Krahvô (Popjes & Popjes 1986), (4), we see that the inclusive is expressed by cu but for the singular as well as the (plural) exclusive the same pronoun is used: wa.

This situation, where the exclusive is syncretic with the first person singular, is relatively common; it has been attested among native American languages, the Papuan languages of New Guinea and there are various incidental examples (Cysouw 2005). Conversely, we do not observe the use of an identical pronoun for the first person singular and the (plural) inclusive, whilst there being a separate special pronoun for the exclusive. It is important to note that in this case, the asymmetry does not seem to be a trend but genuinely unattested.⁸

Finally, another asymmetry, noted in Moskal (2014), is that whilst pronoun suppletion in the context of solely the inclusive is attested in a variety of languages, we do not seem to observe pronoun suppletion only in the context of the exclusive. Moskal (2014) was a preliminary study, based on a limited number of languages. In this paper, I used a database containing 455 languages (Smith 2011) to verify the attested and unattested suppletion patterns. Cast here as 1SG-1EXCL-1INCL triples, I corroborate the findings of Moskal (2014) that although there are five logical patterns, only four out of these are attested: AAA, ABB, ABC and AAB are all observed, but ABA is crucially unattested. As mentioned above, in the Appendix the full list of languages from the database of Smith (2011) that make a clusivity distinction is given, organised according to the category that I assigned them: AAA, ABB, ABC, AAB and ¬ABA. Only one dubious case for ABA is identified, Mangarayi, which is discussed in section 5. The classification ¬ABA encompasses patterns that do not straightforwardly fall into any of the above, but clearly do not constitute an ABA pattern. Chadong (Smith 2011) will serve as an example; the pronominal paradigm is given in (5). The triple je-lje-la:u has two plausible analyses: as either an ABC pattern, where all three forms are phonologically unrelated, or an AAB pattern, in which the base je is shared in the singular and the exclusive, but not the inclusive. Somewhat less plausibly, one could even choose to focus on the shared lateral of the exclusive and inclusive and consider this an ABB pattern. Crucially, however, the Chadong pronouns do not constitute an ABA pattern, and as such they are classified as ¬ABA, without necessarily committing as to whether they should be analysed as ABC or AAB (or ABB).

Chadong also illustrates the difficulties in judging whether suppletion takes place and what kind of pattern a paradigm belongs to. The main issue is that there is no clear consensus as to what should be considered suppletion and what should be considered (morpho-phonological) irregularity (Mel’čuk 1994; Bye 2007; Corbett 2007; Haugen & Siddiqi 2013). The main criterion for suppletion that I used was whether there is an identifiable invariant formative within a given pronominal paradigm, but ultimately the classifications reflect my judgments as to such a formative (see also Smith et al. 2018: §3.2 and §5 for some more discussion of the complications involved in classifying pronominal suppetion patterns). In the Appendix, I have included the full data set of first person pronouns that I have considered (from Smith 2011), and the interested reader is welcome to consult my classifications.

Turning to illustrative analyses of the patterns, first consider a pattern where the base of a free pronominal remains constant, as exemplified by Ayiwo (Smith 2011) in (6). Here, we see that the pronominal base is always realised as ju to which various suffixes are attached in order to derive the relevant pronominal form.⁹ Crucially, when a morpheme remains constant throughout a morphological paradigm this constitutes an AAA pattern.

An ABB pattern is exemplified by the pronominal system of Ura (Smith 2011), given in (7). Here we see that the first person singular form is yau but that there is a clearly identifiable base gi- shared by the exclusive and inclusive plural forms, though each has their own suffix: -m in the exclusive and -s in the inclusive.

Suppletion in the context of both the inclusive and exclusive is also attested. This situation is exemplified in (8) by Dan (Smith 2011), where the bases for 1SG, 1EXCL and 1INCL are phonologically entirely unrelated; as such, the Dan data illustrate an ABC pattern.¹⁰

Turning to the final attested pattern, AAB, consider data from Evenki (Nedjalkov 1997) in (9). Here we observe pronoun suppletion in the context of the inclusive but not in the context of the exclusive. The form for first person singular decomposes into a base b- for first person followed by a suffix -i for the singular. The latter morpheme also surfaces in the second person singular, where s- is the second person base, which is followed by -i in the singular. In the exclusive and second person plural forms, we see that the bases of first and second person are retained, but that the number information is expressed by the suffix -u. In contrast, the plural inclusive form mit is distinct from the first person base used in the singular and exclusive.

It should be noted here that the first person pronouns in Canela-Krahô, in (4) above, do not constitute an AAB pattern (see also Smith et al. 2018). Recall that in Canela-Krahô the singular and the exclusive form are syncretic (wa), and the inclusive is realised by a different pronoun (cu). Whilst one could construe this as an AAB pattern, given that the singular and the exclusive are not suppletive relative to one another, but the inclusive is, this case differs from that seen in Evenki, because there is complete identity of the entire pronominal form between the first two members, whilst in Evenki, there is identity between the base of the first two members, but not the entire pronominal forms. Under an analysis of syncretisms, where the syncretism is not accidental but reflective of literal featural identity, Canela-Krahô should not be seen as AAB, but rather an AB pattern, where the triple has been collapsed into a double as the first two members are non-distinct.¹¹

The final logical configuration, ABA, in which suppletion occurs only in the context of the exclusive is unattested in my survey; three (types of) apparent counterexamples to this claim are discussed in section 5.¹²

In sum, we have seen three instances of an asymmetry between the inclusive and exclusive in languages that make a clusivity distinction. (i) The inclusive seems more often overtly coded by an additional sequence than the exclusive. (ii) The exclusive can be syncretic with first person singular, but the inclusive never is syncretic with first person singular across the exclusive. (iii) Suppletion in the context of the exclusive without also having a suppletive variant in the inclusive is unattested.

3 Suppletion

Recent work on suppletion (Bobaljik 2012; Moskal 2015a; b; Smith et al. 2018, i.a.) has shown that suppletion, rather than being a merely descriptive term, can serve as a detector of morphological structure. Here, I briefly recapitulate the arguments from Bobaljik (2012).

In a study of 73 distinct adjectival cognate triples, Bobaljik shows that not all suppletion patterns in comparative morphology are attested. Specifically, the patterns in Table 1 are attested in languages, whereas the apparently legitimate and a priori conceivable patterns in Table 2 are unattested. Whilst both ABA and AAB are unattested in adjectival suppletion, AAB (greyed out) is attested elsewhere as discussed above, and as such the current paper does not discuss the absence of this pattern in adjectival suppletion any further (but see Bobaljik 2012 how to exclude AAB from adjectival suppletion patterns). The crucial contrast discussed in this paper is the absence of ABA as opposed to the other logically possible patterns.

Bobaljik shows that the absence of ABA patterns is accounted for if we assume (i) the containment hypothesis, and (ii) late insertion. Specifically, the containment hypothesis is formulated in (10):

In effect, (10) proposes that it is a universal property of languages that if there is a superlative in the structure, then there necessarily must be a comparative in the structure. That is, the structure for any given superlative is as in (11), adjusted from Bobaljik (2012).¹³

The second ingredient for ruling out the unattested ABA pattern is the assumption that syntactic structure is the input to morphology, which then has the task of supplementing syntactic structure with phonological material (Vocabulary Insertion, VI) (Distributed Morphology, DM; Halle & Marantz 1993). Crucially, phonological substance is provided post-syntactically (“late insertion”) and occurs cyclically starting from the most deeply embedded element (Bobaljik 2000). In such a framework suppletion is modeled as contextual allomorphy: a feature (set) has a context-free default exponent, but in a more specific context a different exponent takes precedence (Bobaljik 2012). Consider the VI-rules in (12) and (13); these are the rules that are relevant to the abstract item .¹⁴ Whilst (13) has no restrictions with regard to its application, (12) applies in the context of the comparative (and per the containment hypothesis the superlative as well).

Furthermore, per the Elsewhere principle (Kiparsky 1973) the more specific VI-rule in (12) will be preferred over the less specific VI-rule in (13). That is, given that (12) makes reference to the more specific environment of the comparative, it must be employed in that context; the VI-rule in (13) will apply as a default but given the existence of the rule in (12) crucially not in the context of the comparative. The containment hypothesis combined with late insertion thus gives us the tools to derive *ABA: given that the most specific VI-rule must be used, and each superlative must contain a comparative, if the comparative suppletes the superlative necessarily must do so too and it cannot “revert back” to the default. In sum, suppletion data provides crucial evidence for morphological structure; in Bobaljik’s study for the structure of adjectives as shown in (11).

4 Clusivity analysis

Turning back to clusivity-driven suppletion in free pronouns, recall that the following asymmetry was introduced: whilst AAA, ABB, ABC and AAB are attested (Table 3), ABA is unattested (Table 4) (see section 5 for three (types of) apparent counter-examples).

Consider first the notational representations for person in Table 5 (McGinnis 2005; Bobaljik 2008; see also Harley & Ritter 2002; Cysouw 2003; Nevins 2007; Harbour 2011).¹⁵ Following McGinnis (2005) and Bobaljik (2008) in particular, I assume that the privative features in Table 5 are relevant for the representation of person. It should be noted here that the current proposal assumes for expository purposes that third person is featurally represented as the absence of any person features. However, Nevins (2007) argues convincingly that in order to analyse Person-Case Constraint (PCC) effects reference needs to be made to third person, which in Table 5 is impossible since it is the absence of any feature. The current proposal is not necessarily at odds with Nevins’ analysis, and I am not excluding a feature [±participant], which would be additional to the features proposed in Table 5.¹⁶

Taking number into account, this then leads to the representations for first and second person pronouns in languages which make a clusivity distinction as in Table 6; here I assume binary [±singular] for singular–plural systems, following Harbour (2011).

On the assumption that, at least morphologically, [+singular] is unmarked (Bale et al. 2011; Smith et al. 2018), we see that 1SG is the least marked first person value.¹⁷ The representation for a first person singular personal pronoun is given in (14).

With regard to the relative markedness of the inclusive and exclusive, we have seen earlier proposals that the inclusive is more marked than the exclusive (Noyer 1992; Cysouw 2003; Siewierska 2004, i.a.). This then leads to the markedness hierarchy in (15), from 1SG being the least marked to 1INCL being the most marked.

In this paper, I propose that we can derive the clusivity suppletion patterns through containment, i.e. structural markedness. Specifically, on the basis of Table 6, we can then formulate another containment hypothesis, given in (16), which applies to clusivity.¹⁸

Crucially, as can be seen in the representation for exclusive pronouns in (17) and that for inclusive pronouns in (18), the structure of the exclusive ([author, –sg]) is contained within the structure of the inclusive ([author, addressee, –sg]).¹⁹

I assume the structures in (17) and (18) to be the basic morpho-syntactic structures, prior to any morphological operations. I remain agnostic whether portmanteau morphemes are created by pre-VI fusion of morphosyntactic nodes or VI at non-terminal nodes (see Radkevich 2010). That is, the VI-rules below (e.g. in (20a)–(20e)) are ambiguous between a single node carrying multiple features (pre-VI fusion) or VI at a non-terminal node governing multiple features (spelling out a, π or b).

Turning to the attested patterns, we can analyse them as follows. Starting off with the AAA pattern in Ayiwo, repeated from (6) above as (19), the VI-rules for the pronouns would be the ones in (20), with the ones relevant for first person pronouns being a, b, c, and f.²⁰ No suppletion takes place and we observe ju-juŋo-jude with a stable pronominal base as the 1SG-1EXCL-1INCL triple.²¹

Next, we discuss the ABB pattern exemplified here by Ura (Smith 2011), repeated from (7) above as (21). The VI-rules for Ura pronouns are give in (22), with a, d, g and h used in the formation of the first person forms. Crucially, the rule that refers to the non-singular first person context (22d) targets both the exclusive and the inclusive, leading to a single base, with only their number suffixes differing. The rules converge on the Ura triple yau-gim-gis.

Turning to ABC, exemplified by Dan, repeated from (8) above as (23), the VI-rules are as in (24), with a, b and c relevant for first person. We see that in Dan the various pronominal forms have no consistent formatives in common, and each form has its own VI-rule. Relevant still is that the rule in (24c) can apply in the context of both the exclusive and the inclusive, but is only applied in the exclusive, since there is an even more specific rule (24a) for (only) the inclusive context. This results in the Dan triple yi-ma-ko.

The final pattern, AAB was exemplified by Evenki, repeated from (9) above as (25). The VI-rules for the Evenki personal pronouns are listed in (26).²² Focusing on the VI-rules relevant for first person, we see that there is a specific rule for the inclusive, (26a), but that 1EXCL and 1SG both are subject to (26b) and are realised with a pronominal base b-. This leads to the triple bi-bu-mit in Evenki.

Crucially, ABA is ruled out as a possibility due to containment: any suppletive rule making reference to an exclusive environment, [author, –sg], also necessarily covers an inclusive environment, [author, addressee, –sg] (see also Table 6). As such, we can have only a rule referencing [author, –sg], leading to an ABB pattern (see Ura), or an additional even more specific rule referencing [author, addressee, –sg], leading to an ABC pattern (see Dan), but there is no set of VI-rules that would converge on an ABA pattern.²³

To see this, consider the hypothetical ABA triple in (27): ki ‘1SG’, an ‘1EXCL’, kio ‘1INCL’. We would have the VI-rules given in (28): a context-neutral VI-rule expressing author, [D, author] ⇔ ki (28a), and another VI-rule that targets the exclusive context, [D, author, –sg] ⇔ an (28b); for the inclusive environment, we would need an additional VI-rule that realises [addressee] as -o (28c).

In the singular environment, only the context-neutral VI-rule (28a) applies, correctly resulting in ki. In the exclusive environment, both (28a) and (28b) potentially apply, but, per the Elsewhere principle, the more specific rule (28b) wins, correctly predicting that an is the exclusive form. Finally, in the inclusive environment, again both (28a) and (28b) potentially apply, but again the Elsewhere principle leads to the application of the more specific rule (28b) and this, combined with the VI-rule that realises [addressee] (28c), then incorrectly predicts the inclusive form to be ano. This then results in an ABB triple ki-an-ano instead of the ABA triple ki-an-kio. To reiterate, when there is a context-neutral rule for 1SG and a suppletive VI-rule that targets the exclusive context, this rule must also apply in the inclusive context, and, as such, ABA cannot be derived.

In sum, we have seen that the suppletion patterns support the structural representation for first person exclusive and inclusive as in (17) and (18). This account crucially relies on the absence of ABA patterns in the study, and in the next section I discuss purported counter-examples to the claim that ABA is unattested in my sample.

5 Counter-examples

The counter-examples to the absence of ABA patterns in 1SG-1EXCL-1INCL triples can be divided into three types: (i) languages in which the inclusive is transparently composed of a combination of first and second person pronouns (referred to as “dvandva” pronouns here), (ii) a single recurring form occurring in various (cognate) Carib languages, and (iii) the case of Mangarayi.

Turning first to languages that show an inclusive pronoun that transparently contains a first and second person morpheme, consider the pronominal paradigm of Dolakha Newar (Genetti 2007) in Table (29).²⁴ The Dolakha Newar data show a suppletive form in the exclusive, isi, but the inclusive chiji transparently contains the singular form ji to which a morpheme chi is added. Taken in isolation, this would constitute a clear ABA pattern. However, the chi morpheme actually corresponds to the second person singular, as can be seen in (29). As such, the inclusive form in Dolakha Newar can be coined a “dvandva” pronoun, since it is transparently composed of a morpheme for addressee, chi, and a morpheme for author, ji. Indeed, although usually the two person features that the inclusive contains are not expressed separately, from the representation in (18) we see that this simply is a logical possibility that the system provides. As such, I propose that in Dolakha Newar the VI-rules in (30) are active for first and second person pronouns (analysis to be refined below).²⁵

Though intuitively on the right track, these VI-rules make a wrong prediction for the realisation of the inclusive form. Since the exclusive is a subset of the inclusive, the suppletive rule in (30a) for the exclusive environment applies in the exclusive as well as the inclusive. Leaving aside the exact linearisation, the VI-rules in (30) lead us to expect a form isichi, with the suppletive form rather than the first person singular form ji. The suppletive variant is crucially reliant on the [–sg] feature, but the inclusive does not bear any number marking: the form literally only contains the morphemes for singular author and singular addressee. We can thus remedy the situation by positing an impoverishment rule (Noyer 1992) in Dolakha Newar, which deletes the plural in the inclusive (see (31)).²⁶

The impoverishment rule in (31) nullifies the containment relation: the exclusive in Dolakha Newar, composed of [author, –sg] at the point of VI, is no longer a subset of the inclusive in Dolakha Newar, composed of [author, addressee] at the point of VI. Given that containment no longer holds, the 1SG-1EXCL-1INCL paradigm in Dolakha Newar does not constitute a triple for which ABA is predicted to be impossible.

However, we expect that not all languages with dvandva pronouns have an impoverishment rule such as (31). Indeed, Tok Pisin (Verhaar 1995; Smith 2002) also displays a dvandva pronoun of first and second person, as can be seen in (32). Note also that Tok Pisin, in addition to a clusivity distinction, shows a four-way distinction for number, displaying singular, plural, dual and trial. Interestingly, here we see that the non-singular number morpheme -pela, albeit optionally, is realised in the inclusive plural.²⁷

The current account makes a clear prediction for languages which retain the plural in the inclusive, i.e. which lack the impoverishment rule in (31). In such languages, the 1SG-1EXCL-1INCL paradigm constitutes a triple that is subject to *ABA: if the exclusive is suppletive, then the inclusive must be as well. The Tok Pisin data, however, do not show any suppletion, thus being in line with the hypothesis here, but I leave a systematic study into dvandva pronouns for future research.

Turning to the second type of counter-examples to the *ABA generalisation, consider data from Macushí (Abbott 1991) in (33). We see that the first person singular and the inclusive clearly show a base uurî, whilst the exclusive is realised as a phonologically unrelated form anna. Furthermore, the inclusive seems to contain a plural morpheme -(’)nîkon, which can be identified in the second person plural form as well. Thus, Macushí prima facie constitutes a textbook ABA pattern.

This exclusive anna form has clear cognates in various Cariban languages (see Meira 2002), and as such these represent a single data point (see Bobaljik 2012: section 1.3 for discussion). Nonetheless, this is a robustly attested pattern in this family, where the exclusive is realised suppletively, whilst the inclusive shares its base with 1SG. However, there is some reason to be suspicious whether anna is a true pronoun. Firstly, unlike all other first and second pronouns, at least in Macushí, it appears only as a free form and does not have an affixal variant, which as such sets it apart from all other forms. Secondly, it does not seem to control plural agreement on the verb, which seems to hold for all Cariban languages (Meira 2002). Focusing on the latter argument, consider the sentences in (34)–(36) from related Waiwai (Hawkins 1998), where the cognate amna in (36) does not control plural (collective) agreement on verbs, appearing only with singular marking (Hawkins 1998). In (34) and (35), we see that the verb contains a collective suffix -cow when there is a collective argument present, a second plural object (in (34)) or a third person subject (in (35)). However, in (36) the “exclusive” does not result in the collective suffix -cow on the verb.²⁸

The fact that amna does not participate in verbal agreement, I speculate, is indicative of it having a different structure than the other pronouns in the language, and should not be seen as a genuine counter-example to the proposal here. The different structure has the dual effect that it prevents the pronoun from reducing into an affixal form (cf. Cardinaletti & Starke 1999) and makes its phi-information inaccessible to agreement. As such, the Cariban languages do not have a pronoun with the structure in (17), and thus they no longer have 1SG-1EXCL-1INCL triples.

Returning to the Macushí data in Table (33), I posit, without committing to an internal decomposition, that the first and second person pronominal forms contain the features in Table 7; [±augmented] is needed to distinguish between dual and plural.

Note that anna is not listed in Table 7, since it does not represent a true pronoun. In fact, this view that amna has an different structure seems to be supported by Meira (2002), who notes that the pronoun is treated syntactically as if it were third person, and “one wonders if it could have been an old non-possessible noun (cf. e.g. Brazilian Portuguese a gente ‘we’, literally ‘the people’)” (Meria 2002: 257). Though Meira does not chart a particular diachronic path, the quote is certainly suggestive of the opinion that this form is not a true pronoun. I do not have anything further to add to what the structure of amna actually is, which I leave to further research. What is important for now is that there seems to be enough reason to be suspicious of amna being a true pronoun, and we may put it aside, since it should not form part of a legitimate triple pattern.²⁹

Finally, this leaves us with one more problematic case that I encountered in my sample: the case of Mangarayi (Merlan 1989). Although for the most part I have ignored the distinction between singular–plural and minimal–augmented systems (see footnote 4), Mangarayi shows a minimal–augmented number system, which needs to be discussed in some more detail. Whilst the singular–plural distinction is an absolute classification based on “one” vs. “more than one”, the minimal–augmented distinction is relative and based on “logical minimum” vs. “more than the logical minimum” (Corbett 2000). The difference can be captured notationally as follows: Table 8 shows a singular–plural system, and Table 9 shows a minimal–augmented system. The main difference lies in the inclusive: in a singular–plural system it cannot be singular, but in a minimal–augmented system the inclusive can still belong to the minimal category. Translating the notational 1SG-1EXCL-1INCL triple from a singular–plural system to a minimal–augmented system results in 1EXCL.MIN-1EXCL.AUGM-1INCL.AUGM as the correspondent triple.³⁰

With this brief background, consider the data from Mangarayi (Merlan 1989) in (37). Note that Mangarayi makes an additional distinction of unit-augmented, which can be paraphrased as “logical minimum plus one” (cf. dual in singular–plural systems). Mangarayi employs demonstratives to serve as the third person pronouns, and so we will not consider them here.

Following Harbour’s (2016) representation for minimal–augmented–unit-augmented number, I assume the features in Table 10 for minimal–augmented–unit-augmented systems.³¹

As noted above, the corresponding notational triple is 1EXCL.MIN-1EXCL.AUGM-1INCL.AUGM. On the basis of Table 10, we can see that, parallel to the markedness hierarchy for singular–plural systems in (15), we can construct a markedness hierarchy for minimal–augmented systems. On the assumption that [+minimal] is unmarked (cf. [+singular] in section 4), and given that 1EXCL.AUGM is a subset of 1INCL.AUGM, this leads to the markedness hierarchy in (38).³²

Returning to the Mangarayi data in (37), the relevant triple is ŋaya-ŋiḷa-ŋaḷa, where we see that the base is ŋa in the first and third form but ŋi in the second form: a clear instance of an ABA pattern.

However, the problem of the Mangarayi pronoun paradigm runs deeper than the person containment structure proposed here; the real issue is that neither the forms that take ŋa nor the forms that take ŋi as their base form a natural class.³³ Informally, we can already see this in (37): the puzzle is that the ŋa/ŋi distinction tracks neither the vertical axis (person: clusivity) nor the horizontal axes (number: minimal, augmented, unit–augmented). Consider Tables 11 and 12, which list first person pronouns that take ŋa and ŋi as their base, respectively.

Crucially, neither the pronouns that take ŋa as their base, nor the pronouns that take ŋi as their base have a feature in common, other than [author], which however is part of both groups. Consequently, it is not possible at this stage to determine the relevant feature(s) responsible for the irregularity in this pronominal paradigm. That is, irrespective of the containment hypothesis, there is no decomposition available for the Mangarayi pronominal forms that corresponds to the morphological features in Table 10. Solutions include introducing a novel set (and distribution) of features, or considering the pronouns as non-decompositional (effectively, an ABC pattern). Neither of these seem particularly satisfactory, however. Therefore, I note Mangarayi as a potential problem, but pending further investigation, I tentatively suggest that this data point does not provide sufficient evidence of the existence of ABA patterns in clusivity.

6 Markedness

As noted above, the ABA triples considered here are based on the markedness hierarchy in (15) (and the one in (38)). Whilst the parallels to adjectival suppletion (Bobaljik 2012) are clear, it is important to emphasise that only part of the hierarchy in (15) is based on containment. That is, whilst adjectival triples are motivated exclusively by structural markedness (specifically, containment), defined as one element being more marked than another if its structure is more complex, the clusivity triples discussed in this paper are partially motivated by structural markedness (1EXCL < 1INCL) and partially by featural markedness (1SG < 1PL), defined as one feature value of a binary feature being asymmetrically related to its other feature value. In this section, I discuss (some of) the differences between these two types of markedness. Most notably, structural markedness is shown to be more restrictive than featural markedness, since the latter allows for markedness reversals. I will then turn to a discussion of the consequences of this for the representation of clusivity features in the next section.

Both structural and featural markedness as governors of suppletion have been invoked recently in Smith et al. (2018), where we investigate case-driven and number-driven suppletion patterns in nominals.³⁴ Structural markedness, implemented as containment, has been discussed in some detail in section 3, based on adjectival suppletion. Another example is found in case-driven suppletion; Smith et al. (2018) argue in favour of a case hierarchy as in (39), inspired by Blake (1994); Caha (2009) (see also Marantz 1991), where UNMARKED case is nominative or absolutive, DEPENDENT case is accusative or ergative, and OBLIQUE case is usually represented by dative (or another oblique case).

This hierarchy is implemented through structural containment, leading to the representation for case in (40).

Parallel to the triples in adjectives (POS-CMPR-SPRL) and clusivity (1SG-1EXCL-1INCL), we then have the triple UNM-DEP-OBL for case, which is to be read NOM-ACC-OBL in nominative-accusative systems, and ABS-ERG-OBL in ergative-absolutive systems. The suppletive patterns we found in pronouns are given in Table 13: pronouns can lack case-driven suppletion (AAA), as in Lezgian (Haspelmath 1993); pronouns can show one base in the unmarked case and a second base that is shared in the dependent and oblique case (ABB), as in Armenian (Kozintseva 1995); there are cases where all three pronominal forms are fully distinct (ABC), as in Khinalugh (Kibrik & Kodzasov 1990); and examples where a base is shared by the unmarked and dependent cases whilst the form has a distinct base in the oblique (AAB), as in Wardaman (Merlan 1994). No cases of ABA were found.

Crucially, although structural markedness relations can be disrupted and nullify a containment relation, as for instance in the case of Dolakha Newar discussed in section 5, where a feature ([–singular]) was deleted due to an impoverishment rule, structural markedness relations cannot be altered.³⁵

In contrast, number-driven suppletion is not subject to structural but featural markedness according to Smith et al. (2018). Given that three categories are needed, we tested the containment hypothesis in systems that make a three-way number distinction.³⁶ Based on markedness statements such as Greenberg’s (1963) Universal 34 (see (41)), we formulated the markedness hierarchy in (42).

Parallel to the triples in adjectives, case and clusivity, we then have the triple SG-PL-DU for number. The attested patterns in pronouns are reported in Table 14, where we see that pronouns can lack suppletion (AAA), as in Mapuche (Smeets 2008), or that there is one base in the singular and a second base that is shared in the plural and the dual (ABB), as in Sursurunga (Harbour 2014), or there are cases where all three bases are distinct (ABC), as in Tiri (Smith 2011), or instances where a base is shared by the singular and plural whilst the dual has a distinct base (AAB), as in Yagua (Payne & Payne 1990).

However, we found apparent ABA patterns in the lexical nouns in Table 15: in a handful of items, the singular and dual share a form, but the plural shows a suppletive variant. Curiously, number seems to be the only category in which apparent ABA suppletion patterns are observed: adjectives, (pronominal) clusivity and (pronominal) case all support the hypothesis that ABA patterns are unattested.

The reason for the possibility of these apparent ABA patterns will be shown to be due to number triples being based on featural rather than structural markedness in the following way. To start, the hierarchy in (42) is purely descriptive; consider the structural representation of number in Table 16, based on Harbour (2011).³⁷

As Smith et al. (2018) note, although they both contain the feature [±singular], there is no containment relation between the dual (represented in (43)) and the plural (in (44)).

Consequently, the hierarchy in (42) is not based on structural markedness, and ABA patterns are not excluded based on containment relations. Consider the VI-rules in (45) for the suppletive form in Lavukaleve (Terrill 2003); the context of the suppletive rule is [+augmented], which does not target the singular or dual, as can be seen in Table 16, leading to a suppletive pattern, where only the plural has a suppletive base.³⁸

(45)	a.
	b.

However, rather than abandoning all restrictions on number-driven suppletion, we draw on featural markedness in Smith et al. (2018); specifically, we argue that in the context of [–singular], there is cross-linguistic variation as to which value of [±augmented] is marked: either [+augmented] or [–augmented] can be marked. Then, following the idea of Moskal (2014) (which is based on Calabrese 1995; 2005) that marked features but not unmarked features can govern suppletion (see also section 7 below), this leads to the predictions in (46).

The Yagua triple from Table 14 thus reflects a markedness hierarchy as in (46a): SG-PL-DU, and should be read as jiy-jiryéy-sááda. In contrast, the apparent ABA patterns in Table 15 reflect a markedness hierarchy as in (46b): SG-DU-PL, and e.g. the Lavukaleve triple should be read as vo’vou-vo’voul-tulav. Crucially, both Yagua and Lavukaleve (as well as Hopi and Yimas) constitute AAB patterns on this account.

At this stage, it seems an arbitrary choice which value of [±augmented] is (un)marked, other than its suppletive behaviour. As Smith et al. (2018) point out, it is essential for the reasoning not be circular to have an independent determiner of which value is marked. In effect, we argue that featural (and per definition also structural) markedness correlates with overt coding (see also Croft 2003). That is, when a particular feature value is marked, it is overtly (phonologically) expressed by an exponent. To illustrate this, consider the variation in overt coding of plural and dual marking that we found in our study: languages vary in whether the dual is built on top of the plural, or whether the plural is built on top of the dual. For instance, in Manam (Lichtenberk 1983), (47), the dual ŋaradíaru clearly contains the plural ŋaradí, whilst in Panytyima (Smith 2011), (48), the plural nhupalukuru clearly contains the dual nhupalu.

Though neither Manam nor Panytyima display suppletion, they do show cross-linguistic variation as to which value of [±augmented] is marked. Simplifying somewhat for brevity (see Smith et al. 2018 for details), in Manam, the dual is overtly coded on top of the plural, and [–augmented] is marked; in Panytyima, the plural is overtly coded on top of the dual, and [+augmented] is marked.

In sum, Smith et al. (2018) propose that (featural) markedness has a dual role: if [αaugm] is the marked value, then (i) [αaugm] is overtly coded and (ii) [αaugm] can serve as a context for suppletion; in addition, [–αaugm] is unmarked, which means that (i) [–αaugm] is phonologically null and (ii) [–αaugm] cannot serve as a context for suppletion.³⁹ Crucially, in contrast to structural markedness, featural markedness allows for markedness reversals, given that there is cross-linguistic variation as to which value is marked.

Finally, it should be noted here that Smith et al. (2018) argue that structural markedness can be derived from featural markedness through relativisation to only marked features. In effect, the idea is that languages show variation as to whether all or only marked features are visible to VI: the features that are visible to VI (rather than the features present in the morpho-syntactic structure) can then show proper containment. Nonetheless, given that relativisation applies to marked rather than specific values, featural markedness still plays a crucial role. Relevant for present purposes, however, is that there is a distinction between the two types of markedness, with the crucial aspect that reversals are expected in what is referred to here as featural markedness, irrespective of whether this is analysed as featural markedness or as derived structural markedness.

7 Discussion

Here, I turn in some more detail to two aspects discussed in the preceding section, specifically focusing on their consequences for the representation of clusivity features. I will address (i) markedness reversals being a hallmark of featural markedness relations; and (ii) the correlation between featural markedness and overt coding.

We saw that markedness reversals were observed only in number-driven suppletion, and Smith et al. (2018) relate this to cross-linguistic variation as to whether [–augm] or [+augm] is the marked value.⁴⁰ In contrast, in adjectival suppletion, case-driven suppletion and clusivity-driven suppletion no convincing examples of markedness reversals have been found. I propose here that the reason why markedness reversals are not seen in adjectives, case and clusivity, as opposed to number, is that we are dealing with slightly different types of features.⁴¹ Specifically, [±augmented], the feature responsible for the possibility of markedness reversals, is a binary feature, and, as such, either value can be selected as “marked” in a language.⁴² This binary value made it possible to have triples with only two (rather than three) features, and resulted in these triples not (directly) being in a containment (structural markedness) relation. In contrast, the features involved in adjectival suppletion (see (11)) as well as case-driven suppletion (see (40)) have been presented as privative in Bobaljik (2012) and Smith et al. (2018), respectively (though see Müller 2004; Calabrese 2011 among others for binary case features). In this case, the features involved were in a containment relation.⁴³ Thus, the possibility of being in a containment relation is delimited by the type of feature: only privative features can be in a (direct) containment relation. If we then take markedness reversals as the hallmark of featural markedness, we can take the strong hypothesis and posit that, given a large enough sample, we expect that binary features show a markedness reversal. Conversely, the absence of markedness reversals in a large sample, such as the clusivity study here, support an analysis of, at least some, person features to be privative.⁴⁴

This reasoning goes as follows. If the clusivity feature is assumed to be binary, e.g. [–addressee] for the exclusive and [+addressee] for the inclusive, then we expect VI-rules to target either context, thus allowing for a suppletive rule to only target [+addressee, –sg] (an AAB pattern in a 1SG-1EXCL-1INCL triple), as well as allowing for a suppletive rule to only target [–addressee, –sg] (an ABA pattern in a 1SG-1EXCL-1INCL triple). However, the fact that I have found no convincing counter-examples to the absence of ABA patterns in the clusivity data strongly suggests that the clusivity feature, here [addressee], is not [–addressee] or [+addressee], but rather the relation is one of presence vs. absence of the monovalent feature [addressee]. Markedness cannot vary along the same lines as number then, since it would require the literal absence of a feature to be a marked unit, and this is why markedness reversals are not found.

This proposal, however, raises issues for overt coding. The proposal in Smith et al. (2018) to equate markedness with overt coding needs to be addressed. We have seen that both the inclusive (see table (2)) and the exclusive (see table (3)) can be overtly coded, despite the absence of a markedness reversal in the suppletive behaviour. This then, leads me to conclude that overt coding does not always reflect markedness. However, to retain the insights from Smith et al. (2018), I tentatively suggest that overt coding tracks markedness only if complexity or containment are uninformative; that is, overt coding tracks markedness within a single feature. On this proposal overt coding reflects featural but not structural markedness relations. In a binary feature [±singular] it can track either feature value as (un)marked, [+singular] or [–singular]. By way of contrast, in a privative feature [addressee], there is no markedness contrast since there is only one “value”, and, as such, overt coding does not represent markedness. In effect, this means that [+F] or [–F] can be (un)marked, but this does not apply to just [F].

Although this divorces overt coding from structural markedness, this leaves the issue of exponence itself. If structurally unmarked constructions are reflective of the absence of some (privative) feature, then how can they be overtly (phonologically) expressed? Put concretely in the context of this paper, how are we to analyse languages which build the exclusive on top of the inclusive (see also Harbour 2016)? Note that languages that have a unique inclusive morpheme are unproblematic. This morpheme would be the realisation of [addressee] in the context of [author]: e.g. for Itzaj Maya in Table (2), we would have a VI-rule such as [addressee] ⇔ -’ex /__ ] author ]. However, languages that have a unique exclusive morpheme are not that straightforward. Consider the Limbu pronominal paradigm of first and second person in (49) and third person in (50).⁴⁵ Crucially, in the exclusive we see an overt morpheme -ge which does not feature in the rest of the paradigm. The problem here is that since the features of the exclusive ([author, –sg]) are a subset of those of the inclusive ([author, addressee, –sg]), we expect that whilst it is possible that the inclusive has an additional morpheme the exclusive can contain no morphemes that are not also present in the inclusive (see also Harbour 2016).⁴⁶

Before turning to an analysis in line with the assumptions here, let us first see how this is not a problem in a system which assumes a binary feature for the clusivity contrast. In his discussion on overt coding of the exclusive, Harbour (2016) shows that assuming that a binary feature distinguishes between the exclusive and inclusive circumvents this problem. Specifically, in a singular-dual-plural system with a clusivity distinction, he assumes the features in Table 17 to be active; for present purposes, [+author] can be equated with [author] and [±atomic] with [±singular]. [±participant] is the crucial feature for the discussion here, since it refers to the presence or absence of discourse participants, thus disambiguating between the inclusive and the exclusive.⁴⁷

Following the analysis of Limbu in Harbour (2016: 107), an is the realisation of [+author] (with place assimilation from an to aŋ before /g/ in the singular), -chi is the expression of dual and -i of plural.⁴⁸ This leaves us with the morpheme -ge in the exclusive forms, which in a system in which the exclusive and inclusive are distinguished by binary [±participant], is simply the realisation of [–participant]. This then, leads naturally to the attested exclusive and inclusive forms in the first person dual and plural forms. For clarity, the decompositional structure of the exclusive plural and inclusive plural is given in (51a) and (51b), respectively.⁴⁹

As Harbour (2016) notes, however, the analysis raises two questions. Firstly, -ge does not seem to be present in the first person singular, although Harbour (2016), following van Driem (1987), notes that the suffix -ga is similar in shape and could be an exponent of [–participant] after (idiosyncratic) phonological changes. Secondly, -ge also does not surface in any of the third person forms; Harbour (2016) notes two explanations: (i) third person is contextually unspecified for [±participant], or (ii) -ge is the realisation of [–participant], but only in the context of [+author].

Data in (52) from Gumbáiŋgar (Smythe 1948–49) shows an analysis similar to that of Limbu: -gei is the realisation of [–participant], but again it needs to be contextually specified to occur only in the exclusive and not in 1SG or third person (see Harbour 2016: 108 for details).

Although Harbour (2016) can naturally account for the overt coding of the exclusive, the containment relations argued for in this paper no longer hold. That is, in Table 17, there is no longer a subset-superset relation between the exclusive and inclusive. As such, the absence of ABA suppletion identified in section 2 becomes a coincidence: since Harbour (2016) formulates specific VI-rules referring to the exclusive ([–participant]), suppletion for only the first person exclusive becomes a possibility.

If we want to retain the impossibility of ABA patterns on the grounds proposed here, we need to see whether there is a plausible alternative analysis for languages which mark the exclusive with an additional morpheme, such as Limbu, (49)–(50), and Gumbáiŋgar, (52). In the following, I will spell out a possible alternative analysis for the Limbu paradigm, which is in line with privative person features. At the heart of the problem of a specific exclusive morpheme in privative person feature systems lies a lack of morpho-syntactic material that can be converted into phonological exponents. This can be circumvented if we make use of the pronominal base D as an additional locus of VI. Harbour (2016) independently proposes such an analysis, with the main difference being that he treats the category label π as a feature, whilst I use D.

On an analysis of an as the realisation of the pronominal base in the context of first person, rather than first person itself, we see that -ge can then be the realisation of first person ([author]). The default realisation of the pronominal base D is the one that surfaces in second and third person, khɛn.⁵⁰ The VI-rules relevant to the discussion are given in (53), and the decompositional structure of the (plural) exclusive would then be as in (54).

However, this analysis, in which -ge is the realisation of [author], leads to similar questions as we saw in an analysis of -ge as the exponent of [–participant]. Firstly, why does it not seem to surface in the first person singular, and secondly, why does it not surface in the inclusive? Note the similarity to Harbour’s analysis: in both cases, the prediction is that the same exclusive morpheme is present on both the singular as well as the exclusive plural, and in both cases the solutions would be similar in that this marker would only surface in the context of the non-singular. Concretely, here I adopt the analysis in Harbour (2016), following van Driem (1987), where the similarity to -ga is noted, and I assume that the -ge/-ga alternation is the result of (idiosyncratic) phonological changes.

The crucial difference between the two analyses lies in the fact that Harbour (2016) predicts that this morpheme also surfaces in third person, whilst the prediction here is that this morpheme also surfaces in the inclusive. Recall that Harbour (2016) proposed two explanations for the lack of the marker in third person: (i) third person is contextually unspecified for [±participant], or (ii) the marker is only realised in the context of first person. An implementation of these two alternatives is offered below: in (55), an impoverishment rule removes the [±participant] feature in third (and second) person, and in (56) there is a context-sensitive rule, which realises [–participant] in first person (56a) and a default rule, which does not express [–participant] in third person (56b).

In the analysis here, which relies on a privative clusivity feature, the overt realisation of [author] has to be prevented, which can be done by assuming that the features for the inclusive [author, addressee] are realised as a (portmanteau) null morpheme, the rule for which is given in (57).⁵¹ This leads to a decompositional structure of the (plural) inclusive as in (58).

When we compare the two approaches, we seem to be at a bit of an impasse. Both approaches face some problems with overt coding in Limbu: in Harbour’s analysis, the exclusive marker is suspiciously absent in 1SG and third person, whilst in the current analysis, the exclusive marker is suspiciously absent in 1SG and 1INCL. In Harbour’s analysis, this is remedied by either removing [±participant] from the relevant context (55) or having a context-sensitive realisation in the context of first person (56), whilst in the current analysis, this is remedied by have a context-sensitive null portmanteau realisation of [author, addressee] (57). However, in the account proposed here, we additionally made crucial use of D to increase the number of loci for VI. That is, even though, at least prima facie, the morpheme an seems to represent [author], here it is analysed as the realisation of D. Thus, if we would only take into account overt coding, this seems to tip the scales in favour of Harbour’s account.

Nonetheless, Harbour cannot readily account for the absence of ABA patterns in 1SG-1EXCL-1INCL triples identified in section 2. Thus, Harbour’s (2016) conclusion that “the issue [of exclusive marking] is not one of descriptive adequacy, but one of insight” becomes somewhat more complicated. Since both approaches achieve descriptive adequacy in terms of overt coding and suppletive behaviour, they are equal in that respect. With regard to insight, however, they both seem to be only partially successful: binary features make for more straightforward overt coding analyses yet miss the *ABA generalisation identified here; privative features easily explain the absence of ABA patterns, but require a less intuitive overt coding analysis of languages that mark the exclusive with an additional morpheme.

As a final note, I return to the alternative that lies somewhat in-between. As noted above, in previous work (Moskal 2014), I pursued the hypothesis that clusivity-driven suppletion is conditioned by featural rather than structural markedness. According to Calabrese (1995; 2005), phonological processes can be sensitive to all values, marked values or contrastive values of a feature.⁵² Focusing primarily on the clusivity distinction in first person, I proposed that the inclusive was marked, whilst the exclusive was unmarked; applying Calabrese’s insight then leads to either marked features governing suppletion, or both unmarked and marked features governing suppletion, but unmarked features alone crucially cannot govern suppletion. That is, when only marked features govern suppletion, the (marked) 1INCL is suppletive but the (unmarked) 1SG and 1EXCL are not (AAB). When both unmarked and marked features can govern suppletion, both (marked) 1INCL and (unmarked) 1EXCL are suppletive (ABB, ABC). Crucially, the lack of ABA is derived, since this configuration would require only unmarked features to be able to govern suppletion.

The purpose of this paper is not to so much to argue against this previous analysis, but rather to sharpen it. Here I presented an approach based on structural markedness, specifically, containment, which is consistent with a wider body of work that suggests *ABA is ruled out by containment (Bobaljik 2012; Smith et al. 2018), and explored its consequences for the nature of feature(s) involved in clusivity. An approach based on structural markedness is more restricted, since the features need to be privative in order to retain a containment relation, whilst this is not necessary in a featural markedness analysis, where markedness can be evaluated at the level of a feature value (as in the case of number-driven suppletion in Smith et al. 2018). However, we have also seen that a structural markedness account of clusivity runs into problems with overt coding, and a featural markedness account should perhaps not be so quickly dismissed.

What is required at this stage is an in-depth study of languages in which the exclusive is marked by an additional morpheme. Assuming a privative clusivity feature, the task would be to see whether in these languages the exclusive marker can be analysed as a realisation of [author]. Were we to assume a binary clusivity feature à la Harbour (2016), it would be well worth combining the insights of Smith et al. (2018) discussed in section 6; specifically, if markedness indeed correlates with overt coding, then in languages in which the exclusive (i.e. [–participant]) is marked, then we predict that the exclusive should be able to supplete alone, whereas the inclusive should not be able to do so.

8 Conclusion

In this paper, I have shown that in a survey of 233 languages, based on Smith (2011), that show a clusivity distinction, in 1SG-1EXCL-1INCL triples there are no convincing ABA suppletion patterns, where the exclusive suppletes to the exclusion of the inclusive. This I argued to be a consequence of the containment hypothesis applied to clusivity: the inclusive always properly contains the exclusive. Furthermore, based on the robustness of this pattern, this supports a privative feature representation of the feature involved in clusivity, which, as such, allows for only one value of clusivity to be the most marked: the inclusive. However, it should be stressed here that the current paper exclusively investigated free pronominal forms, and did not take into account affixal markers of clusivity.

This was contrasted with number-driven suppletion in Smith et al. (2018), where we observed that there is variation as to which value of [±augmented], plural ([+augm]) or dual ([–augm]), is the most marked, thus leading to two possible markedness hierarchies: SG-PL-DU as well as SG-DU-PL. The fact that this kind of reversal was not seen in clusivity I suggested motivates a view of person features as privative, rather than binary.

If this is on the right track, then markedness reversals would be indicative of binary features, whilst their absence would indicate privative features. This would allow an additional evaluation tool to disambiguate between theories that assume binary person features, such as [±author] and those assuming privative [author].

Throughout this paper I have also been using various different conceptions of the term markedness. Proper containment has been taken to be an expression of structural markedness, as in adjectival suppletion, case suppletion and the relation between the exclusive and the inclusive. In addition, I have discussed featural containment, as in number, which has the crucial characteristic of allowing for markedness reversals.

Finally, I have discussed the relation between overt coding and markedness. Smith et al. (2018) propose to correlate the two based on number-driven suppletion and containment relations in number. However, this was shown to be more complicated for the clusivity data investigated here; rather, I suggested that overt coding reflects featural but not structural markedness relations.

To conclude, we have seen that clusivity-driven suppletion supports the strong position of clusivity as best represented by structural markedness, but that future research, in particular of languages that mark the exclusive with an additional marker, will need to show whether we can retain this position.

Abbreviations

AUGM = augmented, COLL = collective, CMPR = comparative, COMPL = completive, DU = dual, EXCL = exclusive, INCL = inclusive, INP = involved mode of non-past tense, MIN = minimal, O = object, POS = positive, PL = plural, PRO = pronoun, S = subject, SG = singular, SF = stem formative, SPRL = superlative, TP = today past tense, U-AUGM = unit-augmented

Acknowledgements

I am grateful to Peter W. Smith, Jonathan Bobaljik, Caroline Féry, Guido Vanden Wyngaerd, and three anonymous reviewers for their valuable comments on an earlier version of this paper. I would also like to thank the audiences at SinFonIJa 9 and GLOW 40 for their feedback, and in particular Anna Pressler for her assistance with the typological classifications.

Competing Interests

The author has no competing interests to declare.