A very unpredictable ‘person’: A corpus-based approach to suppletion in West Polesian

Full Text

1. Introduction

Bybee (1985: 91) defined suppletion or suppletive paradigms as “inflectional paradigms that have forms built on two or more stems that are etymologically from different sources.” I give more detail on suppletion later on, but as a more familiar example we can think of the English verb to go in the past simple > went (*go-ed; as in jump > jump-ed). Both stems are phonologically very different, furthermore, there is evidence that etymologically they derive from two different roots (OED 2018). Hence, go and went hold a suppletive relation; that is to say, their correlation is semantic, rather than formal (phonological).

West Polesian is a little-known East Slavic variety spoken between south-western Belarus, north-western Ukraine and a small fraction of eastern Poland. The speech community lives in an area with a swampy topography that is frequently flooded in spring. As a result, speakers have been highly isolated from outside groups for centuries, as well as having heavily limited mobility within this region. The isolation has contributed to the preservation of some older Slavic cultural and linguistic features, whilst it has also set the ground for innovations with respect to the East Slavic family. Nowadays the community is increasingly exposed to the surrounding standardized and closely related Slavic varieties (i.e. Belarusian, Russian, Ukrainian and Polish). The language contact resulting from media and education is putting pressure on West Polesian grammar. Nevertheless, this is not the only source of contact and pressure. Since the 1980s the government has been draining the marshes, partly, in order to build roads. This has led to a massive emigration of the younger population to the cities, where, in the case of Belarus, Russian is spoken. As a result, West Polesian grammar can often feel like a crossroad of four main Slavic varieties.

Once I advanced on the transcription of the recordings of the West Polesian corpus, I realized that the noun ‘person’ displayed a peculiar behavior, which differed from what I knew from its Belarusian, Russian or Polish cognates. I was aware of the alternation between the stems ʧolovɪk- and lʲud-, in West Polesian. Yet, further on, I noticed that a third stem, duʃ-, was another juggling ball of the suppletive paradigm of the noun ‘person’:

At first glance, it is tempting to assume that these forms are independent lexical entries (synonyms), as traditionally overabundance (Thornton 2019) has been despised. However, a further cross-Slavic corpus survey showed that duʃ- had cognates in other Slavic languages as a suppletive stem of ‘person’ (see the cross-Slavic survey in (§4.1.)). Having a three-stem inventory, (instead of two) like the vast majority of Slavic languages, would make this phenomenon exciting enough to be studied. But there is an even more interesting twist; despite the inventory of suppletive stems being larger than normal their distribution is very heterogeneous across speakers (and even within their own idiolect) when comparing their use in free texts. That is to say, suppletion makes morphology more complex to learn and retrieve (Bybee 1985), so the more stems involved the more regularity we would expect to compensate it, but I will refute this hypothesis in this article.

Numeral Phrases (NumPs) and government are closely related to the stem alternations of the nouns ‘year’ and ‘person’ in Slavic. For this reason, I start with a short overview of their morphosyntax and outline the particularities of Slavic and specifically West Polesian (where there are dedicated adnumerative forms) (§2) in order to frame this study. Second, I introduce the study, present the methodology (§3) and the results from a cross-Slavic survey of the noun ‘person’ in combination with NumPs (§4.1). This shows that the three stems discussed here have cognates in other Slavic varieties. Third, I present some ‘ideal paradigms’ of the three stems involved in West Polesian (§4.2), admitting that these paradigms are frequently combined and mixed by the speakers. Furthermore, the stems ʧolo'vɪk- and duʃ- present complications for the analysis, as homophonous forms exist with full paradigms (§4.2.1). Fourth, based on Bortnik (1979) and Chumakina et al. (2004)
I propose some putative conditions in order to prove whether the choice of one stem over another is restricted and/or motivated by these (§4.3). Moreover, I prove that sociological factors are not automatic predictors of the use of one form or the other, either (§4.3.3). Finally, I present a summary and conclusions extracted in the light of the inconsistencies between speakers and the peculiar sociolinguistic setting of Western Polesie (§5).

2. Some remarks on Numeral Phrases in Slavic and the peculiarities of WP

Most of the suppletion in the paradigm of ‘person’ in West Polesian and Slavic happens around the cells used with Numeral Phrases (NumPs) and quantification. The syntax of NumPs and quantification is very complex in Slavic. The topic has been widely discussed in the literature (and is still being debated). For more complete descriptions and hypotheses from colleagues belonging to different schools I shall refer to the following: Akiner 1983, Babby 1987, Corbett 1983, Franks 1995, Kim 2009, Madariaga & Igartua 2017, Mel´čuk 1985, Nesset 2019, Nesset & Nordrum 2019, Pereltsvaig 2013, Viellard 2011 and Žolobov 2003, among others. Hence, I do not intend to add anything to this topic, but rather I shall mention some of the peculiarities of West Polesian.

Common Slavic had a dual number that has been lost or heavily eroded in most contemporary Slavic languages with the exceptions of Slovene and Upper and Lower Sorbian. In Common Slavic the numeral ‘one’ governed singular; ‘two’, dual; ‘three’ and ‘four’ nom pl; and higher numerals
gen pl (Akiner 1983, Žolobov 2003). Due to some phonological and morphosyntactic changes, the numerals ‘two’, ‘three’ and ‘four’ (henceforth, lower numerals) ended up merging in most Slavic languages (concerning morphosyntax). West Polesian has developed a dedicated form for nouns governed by lower numerals called adnumerative (adnm) or numerative (Mel´čuk 1985, Nurmio & Willis 2017, Roncero 2021, Žolobov 2003). In West Polesian, this adnumerative (henceforth, adnm) form is often in competition with what (at least phonologically) resembles nom pl and gen sg (4).

Moreover, at least some West Polesian nouns seem to also have a special noun form when governed by a higher numeral (i.e. different from the regular gen pl), which I call greater adnumerative (gradnm) (5).¹

Even if forms like tʃoloˈvɪk were only gen.pl, the solution is to propose that another ‘non-canonical phenomenon’ known as heteroclisis is behind this (Stump 2006). Heteroclisis consists of using two inflectional forms belonging to two different inflectional classes. In this case, the tʃoloˈvɪk (and similar forms for ‘year’ e.g. pjet ɦod, instead of the regular ɦoˈdɪʋ ‘five years’, which is also attested) would have a gen.pl proper of inflectional class i,² unlike the singular sub-paradigm, which belongs to class ii. Note that in West Polesian (as in East Slavic) the nominal paradigms of classes i and ii are only distinct from each on the acc/gen cell(s) of the plural sub-paradigm.

3. Methodology and some representation remarks

In order to make sure that we are not dealing with overlapping synonyms, I have undertaken a cross-Slavic survey to see the behavior of the noun ‘person’ (Table 1; Table 2; Table 3). The materials for the cross-Slavic survey are based on ParaSol, a Corpus of Slavic and Other Languages (Waldenfels & Meyer 2011).

The corpus from which the West Polesian examples in this article have been extracted is the result of eight months of my own fieldwork in the region of Brest (Belarus).³ More than fifty speakers took part in the project, contributing different types of data. In order to protect the anonymity of the participants who did not want to disclose their identities each was assigned a code. The first two/three letters of the speaker code indicate the village where they come from. I will refer to this information further on for comparison. It should be noted that all the examples used in this paper have been exclusively taken from free texts, which have given 245 tokens for ‘person’. For the sake of transparency and data replicability, the full list of tokens can be found in the Supplementary Materials.⁴

The examples in this paper are transcribed according to IPA conventions (except for capitals in proper names), based on my ongoing (impressionistic) analysis of West Polesian phonology. I have tried to respect all the differences in the pronunciation, which means that sometimes stems appear as palatalized (ʧʲelovjek-, lʲud-, as in Contemporary Standard Russian – CSR); or most often unpalatalized (which is more in line with the general phonological rules of West Polesian) and the vocalism is not very consistent.

West Polesian stress is dynamic and has a lexical function. In West Polesian many nominal (and less so verbal) paradigms have mobile stress which helps to disambiguate otherwise segmentally syncretic forms (much more than in Belarusian and Russian); e.g. (B6) [gen sg] kaˈnavɪ vs. [nom pl] kanaˈvɪ ‘channel(s)’; (Z4) [gen sg]ˈpɪsni vs. [nom pl] pɪsˈni ‘song(s)’. Thus, given that in West Polesian case/number marking is more clearly dependent in both stress and suffixation than in other Slavic languages, I do not provide morphological segmentation for nouns. Whenever there are differences regarding the position of the stress between speakers or even the same idiolect I have respected these. Unless specifically glossed, all numerals are cardinals.

4. The noun ‘person’

Under this heading, firstly, (§4.1) I present the noun ‘person’ and its behavior with numeral phrases across the Slavic family distinguishing synonyms from pure suppletive stems. Secondly, I present West Polesian paradigms for ‘person’ to which in an ideal instance speakers would adhere (§4.2); also mentioning some of the interferences derived from homophony (§4.2.1). Thirdly, I propose several putative conditions which could affect the choice of one stem or another, starting with morphosyntax (§4.3.1) and (§4.3.2), and moving to sociolinguistics (§4.3.3).

4.1. The noun ‘person’ across the Slavic family

I ran a cross-Slavic survey of the noun ‘person’ using the corpus by Waldenfels & Meyer (2011). Visualizing the suppletive stems and the available distributions for these across the Slavic family has been enlightening for this study. The cells most affected by suppletion are the ones that are most often employed for numerals (i.e. adnumerative (adnm), genitive plural/greater adnumerative (gradnm) or similar). That is why, in the following table, I pay special attention to the type of numeral each form appears with in direct cases.

Here is a list of the languages included in the survey, classified according to their subfamilies:

South Slavic[5]: Bulgarian (BG); Croatian (HR); Macedonian (MKD); Serbian (SRB); Slovene (SVO).

West Slavic: Czech (CZ); Polish (POL); Slovak (SK); Upper Sorbian (US).

East Slavic: (Standard) Belarusian (BLM); Contemporary Standard Russian (CSR); (Standard) Ukrainian (ULM).

Since the results of this survey are quite heterogeneous within each sub-family, I present them according to their genetic/areal affiliation rather than sorting them by the stems. Where there have been many forms, I have stressed in bold the most common or dominant form. The areas in grey indicate the absence of data or results.⁶

Table 1. South Slavic

Table 2. West Slavic

Table 3. East Slavic

Due to time limitations, I focus only on suppletion in contexts where there are numerals; i.e. where dedicated adnumerative forms are used (cf. Roncero 2021),¹⁰ and nom sg – nom pl, in order to see the contrast.

Whilst the noun ‘person’ in West and East Slavic is fairly unexciting, there is a very rich variation in the South Slavic branch. Other synonyms have been interfering with corpus results. The forms osoba (and the like) and lica (and the like) are present in many Slavic languages alongside the more established forms (človek, ljudi, duši and the like). In most languages I have found enough evidence for stating that such forms are not additional stems, but just synonyms. For example, the following sentence would not be allowed in Slovak if the stems were suppletive (i.e. two suppletive stems in a disjunctive) “<...> obyčajní ľudia či osoby nevedomé” ‘<...> ordinary people or unknown persons’ (Waldenfels & Meyer 2011: 77–439)). After refining the results of the corpus, we can see that the suppletive stems under discussion (for West Polesian ‘person’) are present in other Slavic languages, particularly in the South Slavic branch. However, it is important to remark that the stem duʃ- does not appear elsewhere in (contemporary) East or West Slavic, so this is probably an archaic feature of West Polesian, rather than a contact-induced loan.

4.2. West Polesian suppletion patterns for ‘person’

Based on observations from the corpus, these are the underlying paradigms of the three stems for ‘person’. Note that for lud- many speakers produce it as [lʲud-].¹¹ Nonetheless, for the sake of consistency (and to focus exclusively on suppletion), I have kept the non-palatalized form, in the paradigm below (Tables 4 a, b, c). Thus, I represent the cells containing those stems following a color code: ʧolovɪk-, red; lud-, blue; duʃ-, yellow. Given that the status of the greater adnumerative is arguable, I have represented such forms as gen.pl.

Tables 4 a, b, c Available stems for the noun ‘person’ in West Polesian

The reader may have noticed that what I have included here (Tables 4 a, b, c) as the vocative (singular) form (ʧolo'vɪʧe!) can be either a derived form (with an appreciative), or a non-productive type of stem alternation. In favor of the second position, it must be said that the alternation between -k and -ʧ is almost certainly related to a phonological process from the Common Slavic era known as the “First Palatalization” (Shevelov 1964). This is certainly an old remnant of that, especially considering that the vocative is a marginal case value and perhaps more susceptible to retain older forms (see the discussion on cases in: Corbett 2008, Daniel & Spencer 2009, Zaliznjak 2002). This also happens with some of the Belarusian and Russian remnants of the older vocative; e.g. (CSR) [nom sg] Bog > [voc sg] 'Bože! ‘oh, God!’.

4.2.1. Homophony and ambiguity

One of the biggest challenges for the analysis of the corpus has been the ambiguity caused by the homophony of the stems ʧolo'vɪk- and duʃ-. The latter derives from duʃa ‘soul’ (most likely motivated by metonymy). The form duʃ(a) can be also found unbound, but it denotes literally ‘soul (i.e. spirit)’.

Corbett (2007) and Vanhove (2017) describe a similar problem with a few CSR suppletive nouns such as rebënok-deti ‘child’. The noun rebënok is not available for the plural; and so, for the rest of the forms, it uses the stem det-; i.e. [nom pl] deti; but not *rebënki. The noun ditjo ‘child’ exists on its own, but is rather archaic and restricted to the literature. As a result, the most ‘direct’ or semantically regular [nom pl] of rebënok is deti.¹³

In short, the fact that duʃa can stand on its own and has a meaning closely related to ‘person’, does not impair part of its paradigm from being used as a suppletive form of ‘person’ (rather than it being a mere synonym). Moreover, we have the cross-Slavic survey (Table 1, Table 2 and Table 3) as evidence of this form being employed as a suppletive form of the adnumerative forms, most remarkably in Bulgarian (Table 1). Consequently, from now on, whenever there is a possibility of confusion, I will refer to duʃa¹ as one of the possible suppletive stems of ‘person’; and duʃa², as a proper non-defective noun, meaning ‘soul’.

The forms of ʧolovɪk- are far more complex. There is a continuous overlap of forms. Nevertheless, the noun ʧolovɪk¹ ’person’ is, at least, homophonous with ʧolovɪk² ‘man, husband’, which I am also going to distinguish with supra-indexes for the sake of clarity.

The evidence for this distinction is that in some varieties these nouns inflect differently in the cell of the adnumerative. When it means ‘man or husband’ it has a prototypical adnumerative and gen pl/gradnm of inflectional class II, and syntactically it behaves as a plural noun, as we can see in the examples (7) – (10) (note that (8) and (9) are part of the same text).

I reiterate that such a distinction is not clear-cut and that for many varieties both lexemes inflect the same way.

4.3. Conditions for suppletion for the noun ‘person’

Suppletion makes inflectional paradigms more complex; and thus, less predictable and demanding more memorization effort (Matthews 1991). Ideally, we would expect the speakers to adhere to one (or maximum two) of the afore-mentioned paradigms (Tables 4 a, b, c), but I show how this expectation is far from being met.

In the previous sub-sections, particularly in §4.1, I have demonstrated that the stems involved in this suppletion riddle, were true suppletive stems and not simply synonyms. Now, we could state that there is pure overabundance (Thornton 2011, 2013, 2019), i.e. that two (or more) inflectional forms share the same cell and that they can be used interchangeably in any context (e.g. English burned/burnt). However, this is traditionally seen as a result of an inaccurate or superficial analysis by many linguists. In order to argue for this, I have decided to test the stems in different syntactic contexts or conditions to see if they restrict the use of certain stems. The putative conditions are based on the ones set by Bortnik (1979),¹⁴ for the Russian cognates of the suppletive nouns ‘year’ and ‘person’. In addition to these, I propose additional putative conditions based on observations of the behavior of numeral phrases (including quantifiers and various types of numerals)¹⁵ in the entire corpus of West Polesian.

4.3.1. Unbound or non-governed ‘person’

Even if most of the instances of ‘person’ analyzed here are related to quantification, let us start by looking at contexts in which the noun does not appear governed by a quantifier or NumP. The West Polesian corpus survey shows that all the three stems under discussion are possible even though the suppletion patterns vary from one speaker to another. In any case, as far as data in the corpus can attest, the noun ‘person’ uses at least two different stems in every speaker’s variety.¹⁶

As for the unbound forms, the stem duʃ- is certainly dispreferred, if not ungrammatical. This suggests that, most likely, it replicates the behavior of its cognate in Bulgarian and Macedonian. It is possible to find the form duʃa² unbound, but as I have said, it literally denotes ‘soul; spirit’. Not surprisingly, it seems impossible to say something like ‘nasty people’ using duʃ-¹ (*biztolkovɪ duʃɪ?), but I have found examples of the other two in the corpus. See examples (13) – (15):

In any case, I have only found two instances of unbound duʃa¹ in the corpus (16), (17), and both of them doubtful.

In the first example (16), the stem duʃ- is not governed by a lower or higher numeral, the preceding constituent is a quasi-adjectival numeral (see Corbett 1978). The second example (17) comes from another elderly speaker, whom I have recorded using the stem duʃ- with numerals as a suppletive form of ‘person’. However, it may be that the speaker really had in mind its primary meaning ‘soul’ when she produced this sentence.

4.3.2. Bound forms

Most of the conditions which could have an effect on the distribution of the stems are those syntactic contexts in which ‘person’ is governed.

4.3.2.1. Bound to a cardinal numeral

When the noun ‘person’ is governed by a cardinal numeral all three stems can appear.

4.3.2.2. With question words (q) and quantifiers

When a question word (q) or a quantifier is governing the NP, I have documented the stems ʧolovɪk- and lud-, and so far, I have not found the stem duʃ- used for this. Nevertheless, I do not see strong reasons why it could not be employed in these contexts as well.

4.3.2.3. With the adjective ‘all’

With the adjective vsi ‘all’ only the stem lud- has been attested, and it seems unlikely that ʧolovɪk- or duʃ- would be allowed in such a context.

4.3.2.4. Before the numeral

In West Polesian when a cardinal numeral is used after the noun it determines, it indicates that the quantity is approximate (rather than exact). Such a pattern is not only commonly shared by other Slavonic languages, but according to Greenberg’s 44^th Generalization (1978: 284), it is a frequent cross-linguistic phenomenon. In the examples Ошибка! Источник ссылки не найден. and (26), both produced by B18 within the same story, we can see that in Ошибка! Источник ссылки не найден., the main character had collected exactly two kilos [of berries] (B18 referred to a specific amount), whilst in (26) the character is negotiating with someone who looks at his bucket of berries and calculates an approximate quantity.

In these instances the NP stands in gen pl (or gr. adnumerative), although the numeral may be a lower one. Concerning suppletion, so far, I have only documented ʧolovɪk- and duʃ-, but not lud-. This restriction is also shared by Russian, so there seems to be evidence to believe that the stem lud- is ungrammatical in this type of construction.

4.3.2.5. Approximate quantity

When an approximate quantity (rather than an accurate cardinal form) is used, the stem lud- is not permitted.

4.3.2.6. Perception of quantity

I have found instances of the stems ʧolovɪk- and lud-, with larger and smaller numbers, but not duʃ-, probably because there are not many instances of the stem duʃ- in the corpus. However, there is no evidence that the choice of the stems is related to any pragmatic/semantic factors; i.e. whether the speaker perceives the number of people as small (and thus, easily individuated) or large. In (33) -(35) I show some examples of all the stems being used with large amounts.

4.3.3. Sociolinguistic remarks

The West Polesian speech community lives in a marshy area which has been very difficult to access until recently, keeping it isolated for centuries. Seeing as West Polesian is a non-standardized and minoritized variety suddenly exposed to four standardized Slavic varieties (Polish, CSR, BLM and ULM), language change is accelerating and it is not uncommon to find loans (both lexical and grammatical) from its neighbors. One may be tempted to argue that language contact is responsible for all the irregularities in the paradigms. However, in this case, this is not an entirely plausible answer. As I have shown in §4.1, none of the four surrounding Slavic varieties have the stem duʃ- in their inventories (synchronically).¹⁸ The fact that the stem is present in other Slavic varieties (e.g. Bulgarian), with which the community has not had contact for centuries confirms that this is rather a Common-Slavic peculiarity that most contemporary West and East Slavic varieties have lost.

Having said this, I have only been able to document the stem duʃa (when it is used as a suppletive form of the noun ‘person’) in the speech of the older generation (born before 1950): B1, T1, TL1, Z4 and Z10, who also happen to be all female, except for T1. Nevertheless, my intuition is that the fact that only women appear to use this form in my corpus is either accidental or motivated by an unbalanced gender sample,¹⁹ especially concerning the oldest speakers. Hence, the alternation between duʃ- and the other two stems seems more strongly conditioned by age. I should remark that age is often strongly related to bilingualism/exposure to other Slavic varieties, particularly Russian. On the one hand, children (particularly women) who were of school age during World War II often missed several grades (at the time, most received only four years of schooling, but many women only completed one or two). Moreover, people who attended school before World War II (under Polish rule) received their education in Polish (and thus often can be illiterate in Russian and Belarusian).²⁰ Younger people tend to have received more schooling, which since WWII has been conducted in CSR or BLM. Moreover, younger speakers have generally spent or still spend more time in the cities and towns where Russian is the primary language. Hence, the younger the speaker, the most likely they are to be bilingual in Russian, and hence more likely to code-switch.

I have not been able to find any instances of lud- in TL1’s recordings. This could make us suspect that she does not use the stem at all, although it seems very unlikely. I used to have a similar situation with Z4, who used duʃa (as a suppletive root) in one of the sessions, but I have eventually been able to document a token of lud-, in an oblique case (37). Compare (36) to (37) and both produced by Z4.

Thus, we can at least affirm that, although a minority, the three suppletive stems are present in some idiolects. And by looking at their behavior in other Slavic languages, we have more evidence to state that they are not overlapping synonyms, admitting that some stems are more frequent than others.

Finally, considering other sociological and/or sociolinguistic factors such as class/economic status, the use of one form or the other is not marked by that. The reason for this is that all speakers interviewed belonged to the same social class (rural, peasants or blue-collar workers) and that speaking West Polesian is already quite stigmatized in Belarusian Polesie. Thus, people who try to ‘pass’ as a person belonging to a higher social class would most likely try to speak Russian, and less often, Standard Belarusian. This sort of sociolinguistic setting is not uncommon in many speech communities in the world, although, perhaps, the first person to describe a similar situation thoroughly was Nancy Dorian (2010) and her work on Suffolk Gaelic.

5. Summary and conclusions

I have presented the problem of the suppletive noun ‘person’ in West Polesian. I have started by asking whether the alleged three stems belonged to the same paradigm or whether they are synonyms. Evidence from the cross-Slavic survey has shown that these stems also exist in other better documented Slavic varieties as part of their suppletive paradigms. Narrowing down the focus to the West Polesian corpus, I have then applied some putative conditions that would determine the use of one stem or the other. Once these conditions are applied, the resulting picture can be found in (Table 5), which is far from homogeneous. Having said this, there is a handful of patterns that can be extracted considering the putative conditions and the use of the stems by different speakers:

• The noun ‘person’ has a suppletive paradigm in every speaker’s idiolect, which is usually composed of two stems.
• The stem ʧolovɪk- appears in the nom sg of every idiolect, but cannot appear in the nom pl.
• The stem lud- is not allowed with approximate quantities nor in a prenumeral position.

Nevertheless, there is still a vast amount of variation regarding the distribution of the suppletive stems across speakers of the same village (cf. TL1 with TL3-TL4), and even within a single speaker’s idiolect and, hence, overabundance. For this reason, the speakers in (Table 5) are not arranged according to their origin or alphabetical order. Instead, I have decided to group them based on similarities in their suppletive patterns.

Table 5. Distribution of the suppletive stems based on utterances from the most recurrent speakers in the corpus

The results from the study, rather than being very conclusive, open the door to advance in several bigger questions in morphosyntactic typology and variation studies. First, they challenge some of the assumptions concerning suppletion and the need for some regularity supported by some morphosemantic patterns for the sake of learnability (cf. Bybee 1985, Matthews 1991). Second, morphological overabundance has been often related to purely morphological causes. For example, Thornton (2019) argues that the origin of morphological overabundance is to be found in either suppletion or heteroclisis. This also applies to the West Polesian noun ‘person’. Yet, there are some sociolinguistic factors that may have also motivated the situation at hand and which would be overlooked by most traditional theoretical frameworks. Finally, the findings from this paper also add evidence to Dorian (2010) and Thornton’s (2013) claims that sociologically neutral inter- and intra-speaker variation and morphological overabundance are far more common than most traditional linguistic descriptions and theories want to admit. And for that matter, the role of usage-based corpora from understudied languages will increasingly become relevant in this new theoretical and descriptive wave.

Abbreviations

Glossing abbreviations in this paper are based on the Leipzig Glossing Rules (2015 revision), with some additions:

1, 2, 3: first, second, third person (respectively)

acc: accusative

adnm: adnumerative

BG: Bulgarian

BLM: Standard (Literary) Belarusian

coll: collective numeral

CSR: Contemporary Standard Russian

CZ: Czech

dat: dative

dem: demonstrative

det: determiner

f: feminine

fut: future

gen: genitive

iigen: second genitive

gradnm: greater adnumerative

HR: Croatian

imp: imperative

ins: instrumental

ipfv: imperfective

loc: locative

m: masculine

MKD: Macedonian

n: neuter

neg: negation

nom: nominative

NP: noun phrase

NumP: numeral phrase

part: particle

pl: plural

POL: Polish

poss: possessive

prf: perfective

prs: present

pst: past

q: question particle/marker

refl: reflexive

rel: relative

sg: singular

SK: Slovak

SVO: Slovene

ULM: Standard (Literary) Ukrainian

US: Upper Sorbian

voc: vocative

WP: West Polesian

^{1 In Mel´čuk’s (1985: 430–437) terminology this is an “adnumerative plural” but explaining the choice of my term would involve a long discussion deserving, at least, a paper on its own to provide enough context.}

^{2 Traditionally, in Slavic descriptions this is generally referred to, though inaccurately, as feminine (pace Corbett).}

^{3 See §Appendix I for more details on the villages covered.}

^{4 See list of tokens of ‘person’ on the West Polesian corpus at https://doi.org/10.5281/zenodo.5879004}

^{5 In this survey, I distinguish Macedonian (MKD) from Bulgarian (BG) and Serbian (SRB) from Croatian (HR) as I obtained significantly different results for these pairs; however, I did not find any corpus also covering Bosnian and Montenegrin.}

^{6 Although the vast majority of results come from observations from the ParaSol corpus (Waldenfels & Meyer 2011), I had to use Hrvatski Jeyzčni Portal (2006–2021) and Rečnik na bălgarskija ezik (2018) in order to confirm some of the results, which were otherwise inconclusive by mere observations on the corpus.}

^{7 This form is very marginal and all the corpus results point out that, unless an article is used, the stem duši can only be used with quantifiers. Moreover the Dictionary of the Bulgarian Academy (Rečnik na bălgarskia ezik, 2018 (online) also notes that duši must be used with quantifiers.}

^{8 Only one hit in the entire corpus.}

^{9 As in Bulgarian, this form is marginal and can only appear with an article.}

^{10 Regardless of the morphosyntactic strategy they follow, some nouns in adnumerative may have a morphophonologically dedicated form, others may use the regular nom pl form.}

^{11 And for most speakers there is free variation in the non-direct cases between the two realizations.}

^{12 I have only recorded one instance of duʃ- being used (unbound) in the nom pl, where it seems to mean ‘people’ instead of ‘souls’(17).}

^{13 Some people may argue that this noun also has a nominative plural and vocative plural form, based on truncation [nom pl] rebjata!; [voc pl] rebjat! Although these two forms may be originally related to the singular form rebënok, the correlation between the two is not semantically regular. The singular forms ought to be translated as ‘child’; whereas the plural forms are an informal way of addressing teenagers or adults, and so they could be translated as ‘guys’ (see also, Daniel & Spencer 2009).}

^{14 In Chumakina et al. 2004: 294)}

^{15 I tried adding an extra condition: ‘the noun ‘person’ with collective numerals’; but, unfortunately, there are no instances of collective numerals + ‘person’ in the corpus.}

^{16 I.e. it displays a suppletive paradigm in the speech of every participant for whom there is enough data, unlike the suppletive noun ‘year.’}

^{17 Note that the speaker was aware of the suppletive stem duʃ-, but probably in an attempt to be understood by me, an outsider, she tried to simplify the paradigm for me.}

^{18 It was used in Russian and Polish up to the end of the 19th century or the beginning of the October Revolution to refer mainly to servants (see for example Gogol’s Dead Souls), i.e. the use was more nuanced and stylistically marked.}

^{19  I admit there is a gender imbalance in the corpus. From the whole corpus, only fifteen men (27.77%) took part at least in some way, in contrast with thirty-nine women (72.22%). Besides, men’s interventions were considerably shorter than women’s (often participating as ‘people in the back’). This is related to two main factors. First, life expectancy for men is considerably shorter than for women in the area; in fact, a large number of the women I interviewed were widows. Second, in the rural areas where I carried out fieldwork most men under fifty-five work jobs outside of their village (primarily seasonal work in the building industry). Thus, finding men for interviews was more challenging.}

^{20 One of the older speakers who was schooled during the Nazi occupation reported receiving her education in Standard Ukrainian. The remaining people interviewed were schooled in Polish, Russian or Belarusian, if they received any formal education.}

About the authors

Kristian Roncero

Author for correspondence.
Email: kristianroncero@gmail.com
Deutscher Platz 6, 04103 Leipzig, Germany

References