Version classiqueVersion mobile

Dérivation nominale et innovations dans les langues indo‑européennes anciennes

 | 
Alain Blanc
, 
Isabelle Boehm

Première partie. La dérivation depuis l'indo‑européen jusqu'aux langues attestées

Slavic ‘i‑stem adjectives’ and their alleged inflection loss: the derivational prehistory and synchronic status of a category

Marek Majer

Résumé

L’interprétation traditionnelle – explicite ou non – tient que le slave commun a hérité un certain nombre d’adjectifs en ‑i‑ de l’indo‑européen (ou bien d’un ancien dialecte post‑indo‑européen qui aurait créé des adjectifs en ‑i‑ non composés, parallèlement à ce que l’on trouve dans la plupart des autres branches : cf. hitt. palḫ‑i‑ « large », véd. śúc‑i‑ « brillant » etc.) ; cependant, en raison de leur structure anormale ainsi que des évolutions phonologiques à l’intérieur des syllabes finales en slave, ces adjectifs auraient complètement abandonné leurs désinences flexionnelles. En conséquence de cela, ils apparaissent dans les vieilles langues slaves comme indéclinables : cf. v.‑sl. svobodь « libre », različь « diverse », etc., où le ‑ь final invariable réfléchit le thème en *‑i‑ originel. La présente étude fait valoir que de telles formes ne continuent pas directement des adjectifs en ‑i‑ indo‑européens ; elles représentent plutôt l’usage attributif des formes adverbiales qui remontent finalement à l’accusatif singulier du type indo‑européen lat. inermis « non armé », gr. ἄναλκις « sans force », etc. Il n’est donc pas nécessaire de supposer une « perte de flexion » (ce qui serait un développement très étrange en slave commun). Le présent article propose également d’analyser comme parallèles aux développements historiques des adjectifs historiquement indéclinables, certains développements structurellement similaires des langues slaves modernes.

Texte intégral

1. Introduction: Slavic uninflected ‘i‑stem adjectives’

  • 1 I would like to thank the participants of the Rouen conference for their helpful remarks and sugge (...)
  • 2 Note that this fully overlapping distribution makes the Sl. phenomenon quite distinct from seeming (...)

1The oldest Slavic attests uninflected forms in *‑ь otherwise behaving entirely like adjectives:1 they are found in nominal phrases where agreement in number, gender, and case would be expected, in both attributive and predicative position2 (for the data see Diels 1963, p. 191; Vaillant 1948, pp. 127‑128 and 1958, pp. 539‑540; Duridanov et al. 1991, pp. 202‑203). Since the final *‑ь can be mechanically transposed back to PIE *‑i or *‑iC, the forms in question are commonly referred to as i‑stem adjectives (more on which later in this section). Some OCS examples are provided below; example ‎(3) illustrates the contrast between an uninflected adjective (različь ‘diverse’) and a regular, inflected o/āstem adjective (mъnogъ ‘great’) qualifying the same noun.

1. svobodь bǫd‑ete (expected nom.pl.m agreement)
free‑uninfl be‑fut.2pl
‘you will be free’

 

2. m[ъ]nog‑u i različь gněv‑u (expected dat.sg.m agreement)
great‑dat.sg.m and diverse‑uninfl anger‑dat.sg.m
‘to a great and diverse anger’

 

3. rǫc‑ě krъv‑e isplьnь (expected nom.du.f agreement)
hand‑nom.du.f blood‑gen.sg.f full‑uninfl
‘hands full of blood’

 

2Most of the forms belonging to this class are morphologically complex, usually displaying the structure prefix + root + *ь. They are also found as adverbs, i.e. modifying heads other than nominals. Both of these important facts will be discussed further below.

3Uninflected adjectives in *‑ь typically attest derived doublets in *‑ь‑nъ (i.e. *‑i‑no‑ in PIE terms), which are regular o/ā‑stem adjectives and are fully capable of inflection: cf. uninflected *svobodь ‘free’ (OCS svobodь, ORu. svobodь) vs regularly inflected *svobodьnъ, ‑a, ‑o ‘id.’ (OCS svobodьnъ, B/C/M/S slȍbodan, Ukr. svobídnyj, Pol. swobodny) or uninflected *orzličь ‘diverse’ (OCS različь, ORu. različь) vs regularly inflected *orzličьnъ, ‑a, ‑o ‘id.’ (OCS različьnъ, Sln. razlíčən, Ru. razlíčnyj, Cz. rozličný). When the adjectives are used in their form in *‑ь, however, they always remain uninflected; no i‑stem inflection of the type well‑known from nouns – e.g. gen.sg *‑i (< PIE *‑ey‑s), dat.sg ‑i (< PIE *‑ey‑(ey)), gen.pl *‑ьjь (< PIE *‑ey‑ohom vel sim.), dat.pl *‑ьmъ (< PIE *‑i‑bʰ/m‑), etc. – is ever found.

  • 3 For example, traces are extant in certain varieties of Pol. until the 18th century (cf. Jaros 2016 (...)

4These aberrant uninflected adjectives are mostly known from the written corpora of OCS and ORu.; they eventually disappear from later Sl., usually through the prevalence of an inflected derivative of the above type (*‑ь‑nъ) or similar derivational processes. In certain Sl. languages, however, uninflected adjectives in *‑ь persisted as a distinct class for a longer time.3

  • 4 However, as is well‑known, the status of the category in PIE is somewhat elusive, since direct equ (...)

5As was pointed out above, these uninflected forms in *‑ь are routinely treated as a morphological reflex of PIE i‑stem adjectives. In this way, they are identified as the remnants of erstwhile fully inflected, bona fide i‑stem adjectives known from the ancient IE languages (on the basis of which the existence of i‑stem adjectives is inferred for PIE),4 cf. e.g. Hitt. palḫ‑i‑ ‘broad’, Ved. śúc‑i‑ ‘bright’, Gr. τρόφ‑ι‑ς ‘stout, large’, OIr. maith ‘good’ (< *mat‑i‑).

  • 5 If a substantivized neuter of *dʰowgʰ‑i‑ – probably the morphological type of Gr. τρόφις built to (...)

6Note that, in all probability, i‑stem adjectives were also retained as an inflectional class in the closely related Balt.; they appear to be extant in remains in the attested languages (Stang 1966, pp. 259‑261; Petit 1999, pp. 62‑63), cf. e.g. OPr. nom.sg.m arw‑i‑s ‘true’, nom./acc.sg.n arw‑i ‘true’ (adverbialized), OLith. nom.sg.m did‑i‑s, loc.sg.m did‑i‑me ‘big’, Lith. daũg ‘a lot’.5

  • 6 Two PSl. adjectives – *tęžьkъ ‘heavy’ and *gorьkъ ‘burning, bitter’ – have been suspected of conti (...)

7Accordingly, it is customarily assumed that PSl. inherited adjectives in *‑i‑ as well, and that they are in fact still directly attested in the Sl. texts – only without inflection.6

8Put differently, the uninflectability of the forms in *‑ь in the attested Sl. corpus is explained as due to a loss of inflection, typically justified – if at all – as due to the general marginalization of the class and the phonological merger of a subset of the nom.sg and acc.sg endings. Cf. the following representative statements (emphasis in all quotations mine):

Der Übergang der slav. i‑Adjektiva zur Unflektierbarkeit erklärt sich m. E. aus dem Umstand, daß in dieser Sprachgruppe *‑is, *‑in und *‑i lautgesetzlich zusammengefallen sind. (Stang 1939, p. 101)

Ces mots sont d’anciens adjectifs en ‑i‑ qui ont perdu leur flexion parce que le nom. sg. masc. fém. *‑is, l’acc. sg. masc. fém. *‑in et le nom.‑acc. sg. neutre *‑i ont tous abouti à ‑ь. (Stang 1973, p. 77)

Die urspr. adj. i‑Stämme sind indeklinabel geworden. (Vondrák 1924, p. 641)

[C]e sont d’anciens adjectifs en ‑i‑ qui, ne distinguant pas au nominatif‑accusatif singulier le masculin (v. pr. ‑is, ‑in) du neutre (v. pr. ‑i), ni sûrement le féminin du masculin (lit. ‑ė est du type en ‑yo‑, fem. ‑yā‑), ont été considérés comme invariables. (Vaillant 1958, p. 540)

Były to jednak formy już nieodmienne.
However, these forms were not inflected anymore. (Stieber 1979, p. 160)

Лишь небольшая группа прилагательных, утративших формы склонения, сохранила рефлексы тематического *ĭ.
Only a small group of adjectives, which lost their declensional forms, preserved the reflexes of the thematic vowel *ĭ. (Xaburgaev 1986, p. 239)

9This list could easily be expanded; this is the standard description of these forms in the literature.

2. The problem: is a simple loss of inflection plausible?

10However, if taken at face value, a development like the one described in the preceding paragraphs – a diachronic loss of inflection in a specific word‑formation category – should be considered a non‑trivial innovation; it should be carefully inspected.

11Namely, such an interpretation implies that the lexemes belonging here did formerly inflect and that we would in principle expect the same i‑stem endings that are attested in the nouns. Consequently, to reuse the OCS examples ‎(1)‑‎(3) provided above, we could anticipate them to surface as something like:

  • 7 Cf. PSl. substantival i‑stem nom.pl.m *‑ьje, as e.g. in OCS tatie ‘thieves’ (either directly  PIE (...)
4. **svobod‑ie bǫd‑ete
free‑**nom.pl.m be‑fut.2pl
‘you will be free’7

 

  • 8 Cf. PSl. substantival i‑stem dat.sg.m *‑i, as e.g. in OCS pǫti ‘way’ (either < PIE *‑ey‑ey or hapl (...)
5. m[ъ]nog‑u i **različ‑i gněv‑u
great‑dat.sg.m and diverse‑**dat.sg.m anger‑dat.sg.m
‘to a great and diverse anger’8

 

  • 9 Cf. PSl. substantival i‑stem nom.du.f *‑i, as e.g. in *kosti ‘bones’ (< PIE *‑i‑h1; discussion in (...)
6. rǫc‑ě krъv‑e **isplьn‑i
hand‑nom.du.f blood‑gen.sg.f full‑**nom.du.f
‘hands full of blood’9

 

12Instead, however – as the argument apparently goes (cf. especially the statements by Stang or Vaillant quoted in 1 above) – speakers innovated by introducing original nom/acc.sg forms like *svobodь, *orzličь and *ispьlnь into all slots of the paradigm.

13Surprisingly little doubt has been voiced concerning the credibility of such a development. A rare explicit expression of such skepticism (or at least profound surprise) is found e.g. in Meillet 1905, p. 266: “La perte de toute trace des cas autres que ceux qui ont abouti à ‑ь en slave et l’emploi de cette forme sans acception de cas, de genre et de nombre sont des choses très surprenantes”.

14In general, however, the issue is not much discussed; it has attracted far less attention that many “overt” cruces of Sl. historical morphology, such as the reconstruction of particular inflectional endings. Quite tellingly, the recent comprehensive and authoritative handbook of PSl. inflectional morphology (Olander 2015) makes no mention of the loss of inflection in i‑stem adjectives.

  • 10 See, however, fn. 6 on the two adjectives (*tęžьkъ ‘heavy’ and *gorьkъ ‘burning, bitter’) which ha (...)

15In actuality, however, it is scarcely likely that a language with such robust inflectional morphology as (Pre‑)PSl. would have simply given up inflecting an individual class of nominals because of the phonological merger of a subset of endings. Such a phenomenon would be quite unparalleled in the language. Note that the operation of various Auslautgesetze triggered the falling together of quite a number of inflectional endings in PSl., which was generally tolerated (as it happens, the substantival i‑stems provide one locus of profuse phonologically‑induced syncretism, with the ending *‑i coming to signal a whole array of case‑number combinations). To the extent that such situations were not tolerated, responses universally consisted in morphological repair: analogical replacement of single endings, wholesale transfer to more transparent inflectional types, or application of derivational morphology converting a given class into a regular type. This latter mechanism would have been the obvious choice in the case at hand. Note that, as is well‑known, u‑stem adjectives were routinely transferred to the regular o/āstem type via the extension with the suffix *‑kъ (< *‑ko‑). This development essentially led to the unification of the entirety of PSl. adjectival inflectional morphology under the o/āstem class and it is difficult to understand why the i‑stem adjectives should have chosen a wholly different trajectory instead.10

  • 11 Whether such an innovation would have indeed ‘repaired’ the grammar in this context is far from ob (...)

16Conversely, a repair (?)11 strategy entailing the continued use of a given formation with the concomitant loss of inflection (i.e. the extension of a nom.sg form across the full inflectional paradigm of 3 genders, 3 numbers, and 7 cases) – as in the purported case of i‑stem adjectives – cannot be documented for a single other instance in the history of PSl. The question whether such a development would have been plausible is a serious one; it is further discussed in the next section.

  • 12 On definite/long forms – including in the context of the uninflected adjectives in *‑ь – see also (...)

17Finally, it should be noted that – just as it is difficult to justify the supposed loss of inflection in i‑stem adjectives as a natural consequence of the phonological merger of a subset of endings – it is also hardly possible to explain the development as resulting from the “marginal status” of the category. Note that, from the Indo‑European point of view, the inflection of adjectives and substantives was of course essentially identical, and this state was still generally preserved in PSl. (as seen in the o/āstems).12 Had the language “decided” to retain istem adjectives as a separate inflectional class (contrary to the general tendency, cf. above), such adjectives would have had a close counterpart in i‑stem nouns, which were anything but marginal and featured a clear, distinct inflectional pattern.

18To sum up, the tendency to reduce the number of inflectional models of adjectives was evident and the motivation for the very removal of istem adjectives as an inflectional type can be regarded as persuasive. However, this goal would have been easily achieved by the switch to *‑yo/yā(> PSl. *ь/‑ā), *‑i‑ko‑ (PSl. *‑ь‑kъ), or a similar formation. The ostensible response consisting in renouncing inflection and extending the form in *‑ь to all slots of the paradigm – which, far from restoring morphological transparency, would have further obscured the grammar – seems extremely suspicious at best.

3. Loss of inflection: parallels in attested Slavic?

19Now, the argument could be raised that the process of certain originally inflectable lexemes becoming uninflected is indeed at times found in the development of the historical Sl. languages; this fact could potentially be used to corroborate the claim concerning the loss of inflection in (Pre‑)PSl. itself. In reality, however, it is doubtful that any veritable parallels exist: the context of such historically attested changes involves certain components that are absent from the background of the PSl. forms in *‑ь. Two cases are briefly reviewed below.

  • 13 On the IE etymology cf. e.g. ESJS, vol. X, p. 613; Dunkel 2014, p. 111. PSl. *ovъ is usually treat (...)
  • 14 Constructed by the author as standard‑usage counterparts of ‎(10)‑‎(12) below.

20One interesting example is the Pol. demonstrative pronoun ów. In modern Pol., it is only used in formal style and almost exclusively with temporal or discourse reference (hardy ever spatial), mostly in order to avoid the repetition of the conventional demonstrative ten ‘this’. The pronoun possesses a full inflectional paradigm inherited from PSl.: m ów, f owa, n owo (< PSl. *ovъ *ova *ovo,13 cf. OCS ovъ ova ovo ‘that, another’, B/C/M/S òvāj òvā òvō ‘this’, etc.), featuring usual inflection for number, gender, and case. Cf. the following examples:14

7. ow‑a kobiet‑a
this‑nom.sg.f woman‑nom.sg.f
‘this woman’

 

8. ow‑ego człowiek‑a
this‑gen.sg.m man‑gen.sg.m
‘of this man’

 

9. w ow‑ym dn‑iu
on this‑loc.sg.m day‑loc.sg.m
‘on this day’

 

  • 15 In the inflectional morphology of Pol. (and most of modern Sl.), usually acc.sg.m = nom.sg.m in no (...)
  • 16 Examples sourced from Google searches by the author; further examples available (with discussion) (...)

21However, because of the markedly unusual shape of the nom(/acc.)15sg.m ów, which cannot be transparently assigned to any inflectional paradigm, some speakers apparently parse ów as an uninflected element and thus also use it for all other number, gender, and case forms. Cf. the following substandard examples,16 corresponding to ‎(7)‑‎(9) above:

10. #ów kobiet‑a
this‑#uninfl woman‑nom.sg.f
‘this woman’

 

11. #ów człowiek‑a
this‑#uninfl man‑gen.sg.m
‘of this man’

 

12. w #ów dn‑iu
on this‑#uninfl day‑loc.sg.m
‘on this day’

 

22This usage is considered grossly incorrect, attracting reproach in various online services offering guidance on grammar and style (e.g. Sobotka 2006; Malinowski 2016). It is difficult to say how old the phenomenon is, although the fact that it is not mentioned in the entries on ów in major reference works on prescribed Pol. usage – e.g. Markowski 2004, p. 738 – suggests that it is probably fairly recent; needless to say, proper research would be necessary to answer the question definitely. In any case, there can be no doubt that a diachronic loss of inflection has occurred in this lexeme in some substandard varieties of Pol.

  • 17 As discussed at some length in the references provided above.

23That being said, it would be unjustified to claim that the case may serve as a parallel to the development allegedly observed in the PSl. reflexes of PIE i‑stem adjectives. Namely, in Pol., the phenomenon is patently connected with the imperfect command of a prescribed, standardized norm.17 Since the pronoun is generally absent from the vernacular, speakers may – when confronted with the lexeme in formal or stylized language – wrongly extrapolate from nom.(/acc.)sg.m contexts such as ów‑dzień‑ ‘this‑nom.sg.m day‑nom.sg.m’ that ów is an uninflected form (rather than arrive at the correct conclusion that the nom.[/acc.]sg.m is encoded by a zero ending). Thus, the phenomenon is sociolinguistically conditioned and strictly related to a misanalysis of an item proper to a literary standard that one strives to use. Needless to say, the conditions in (Pre‑)PSl. could not have involved anything of this sort. Besides, as noted above, the form in *ь would not have been morphologically opaque at all: the clear model of i‑stem nouns would have been available. Hence, the purported development in (Pre‑)PSl. could scarcely have operated along similar lines as in the case of the Pol. demonstrative ów.

  • 18 In older Sln., the pronoun also occurs in the variant kir; it also partly continues the PSl. relat (...)
  • 19 Examples sourced from Google searches by the author.

24Another type of loss of inflection that is sometimes found in the history of Sl. pertains to items which are involved in processes of grammaticalization (or of an increase in grammatical load) and therefore susceptible to general phonological erosion. Thus, for example, the fully inflected Post‑PSl. relative pronoun *kъjь ‘who, which’ has been reduced in Slovene into an invariable relativizer ki.18 The latter element being uninflectable, the grammatical relations with the relative clause are now expressed via pronominal clitics, as seen in the examples presented below.19 In example ‎(13), no clitic is necessary, since the head noun would also be the subject of the relative clause; in ‎(14) and ‎(15), however, acc.sg.m and dat.sg.f clitics can be seen, obligatorily employed (in lieu of the formerly inflecting relative pronoun) in order to signal the syntactic relationship:

13. študent, ki dela
student‑nom.sg.m rel‑uninfl work‑3sg.prs
‘the student who works’

 

14. študent, ki ga poznam
student‑nom.sg.m rel‑uninfl 3.acc.sg.m know‑1sg.prs
‘the student whom I know’

 

15. ženska, ki ji zaupam
woman‑nom.sg.f rel‑uninfl 3.dat.sg.f trust‑1sg.prs
‘the woman whom I trust’

 

  • 20 Except for the derivation of by‑forms in *‑ь‑nъ (< *‑i‑no‑) etc.; see in 1 above.
  • 21 For some further Sl. examples of this type cf. Willis 2008.

25Once again, this situation is altogether different from the supposed loss of inflection in (Pre‑)PSl. i‑stem adjectives, where no processes such as the ones described above occurred. The lexemes involved are ordinary adjectives with no discernible propensity for grammaticalization. The assumed loss of inflectional capacity in the adjectives themselves is not compensated for by any other morphological means (such as pronominal clitics).20 In conclusion, this type of fall of inflection21 cannot quite account for the PSl. situation.

4. Towards alternatives

26To recapitulate the preceding two sections, it appears that we cannot simply go by Stang’s (1939, p. 102) assertion that “[a]ls sich das Slav. zu einer Sprache entwickelte, wo alle Adjektiva im Nom.‑Akk. ein Genusmerkmal hatten, wurden svobodь im Ausdruck žena svobodь und sugubь im Ausdruck sugubь bogatьstvo als unflektierte Formen aufgefaßt. Dann fing man an, sie auch im Plural zu verwenden […]”. Firstly, the motivation for such a radical change appears far from sufficient. Secondly, the response is quite different from what PSl. (and other older IE languages) usually did; potential parallels in the modern Sl. languages prove to have arisen under markedly different circumstances.

27In view of this, the PSl. uninflected adjectives in *‑ь are in need of an explanation that does not assume the distinctly odd innovation of a simple loss of inflection.

28The obvious alternative, of course, is that the relevant forms never displayed agreement for number, gender, or case in the first place. This would be tantamount to stating that the forms in *‑ь were adverbs to start with; they would have developed adnominal use without ever coming to inflect. The reference to adverbs is not a mere shot in the dark here, of course. On the contrary, adverbs in *‑ь are an important class in Sl. (see e.g. Vaillant 1958, pp. 682‑688; Jelitte 1961, pp. 100‑106). They are, indeed, far more widespread than the uninflected adjectives in *‑ь and constitute a productive type; many are reconstructible for PSl., while numerous others arose in the individual Sl. languages. These adverbs are usually derived from adjectives (e.g. *pravь adv ‘straight ahead’ ← adj *pravъ ‘straight’; *pročь adv ‘far away’ ← adj *prokъ ‘other, remaining’) as well as from prepositional phrases; in fact, they often represent univerbations of the latter (e.g. adv *bezdobь ‘in an untimely fashion’ ← prep *bez ‘without’ + subst *doba ‘[right] time’; adv *osobь ‘separately’ ← prep *o ‘about, concerning’ + pron.refl *sob‑).

  • 22 This example, though widely used to illustrate the type, is somewhat problematic in view of the ev (...)
  • 23 Thus, e.g. the adjectives *pravъ ‘straight’ and *prokъ ‘remaining’ referred to above are compounds (...)
  • 24 The alternative analysis by Meillet (1905, p. 266), according to which the Sl. forms in *‑ь are hi (...)

29Since the great majority of the examples are morphologically complex (prefix + root + *‑ь), the type is routinely treated as an adverbialized nom./acc.sg.n of the well‑known PIE compounding model in which an o‑stem or āstem simplex is shifted to i‑stem form when becoming the second member of a compound (see e.g. Grestenberger 2017; Balles 2006 with ample further literature – prominently works by Schindler, Nussbaum and others): cf. Lat. subst arma ‘weapons’ → adj inermis ‘defenseless’, Gr. subst ἀλκή ‘strength’ → adj ἄναλκις ‘without strength’,22 etc. In Sl., the process was also extended to simplicia, although many simplex adjectives forming the bases of adverbs in *‑ь are historically compounds23 as well.24

30Thus, can the uninflected adjectives in *‑ь simply be explained as adnominally used adverbs in *‑ь?

31This notion is by no means new. The surface similarity of the two classes is obvious and the forms have been routinely treated together from early on (cf. e.g. Miklošič 1868‑1874, pp. 159 and 652; on Miklošič’s analysis see also Grošelj 2007, pp. 123‑124). As a matter of fact, occurrences in texts are often ambiguous between an analysis as an adverb in *‑ь or an uninflected adjective in *‑ь, especially in predicative position (cf. e.g. the discussion on OCS prěprostь ‘simple’ by van Wijk 1931, p. 196). In dictionaries, however, the lemmas are usually separated (e.g. SJS, vol. I, pp. 801‑802: isplъnь1 “adj. indecl. […], plenus” and isplъnь2 “adv. modi […], plene”).

  • 25 Cf. fn. 24 towards the end.

32Although assuming some sort of connection between (the relevant subclasses of)25 adverbs in *ь and the indeclinable adjectives in *‑ь is widespread, the idea that the adjectives might simply represent the adverbs used adnominally – which would readily explain their invariant form and free us from the necessity of explaining the alleged loss of inflection – has been viewed with mistrust across most of the modern literature. Many authoritative sources disavow it unequivocally:

Diese Adjektiva […] sind keine Adverbia auf ‑ь, von denen das Abg. eine ganze Anzahl kennt, sondern wie diese erstarrte Überreste der verloren gegangenen i‑Adjektiva. (Aitzetmüller 1991, p. 130)

[C]e ne sont pas des adverbes qui feraient fonction d’adjectifs, car le fait n’apparaît pas avec des adverbes d’autres types. (Vaillant 1958, p. 540)

33Is this outright dismissal of the idea justified, however? In order to evaluate the solution, it would be useful to investigate the way in which other kinds of indeclinable adjectives arose in the attested history of the individual Sl. languages, or in other IE languages where such a change can be observed diachronically.

5. Parallels for the rise of uninflected adjectives from adverbs

  • 26 For the implication of a good/low/favorable price in the foundational phrase, cf. the semantic dev (...)

34Thus, for example, the basic adjective for ‘cheap’ in standard Slovene is uninflected pocẹ́ni. Historically speaking, it is a univerbation of the prepositional phrase po ceni ‘at a (good)26 price’, i.e. the preposition po followed by the noun cẹ́na ‘price’ in the loc.sg; this analysis is still transparent in the modern language. Nevertheless, the lexeme pocẹ́ni has a full capacity of being used adjectivally in both attributive and predicative role (cf. Greenberg 2006, p. 42). It can occur in any position that ordinarily declinable adjectives can occupy. Where agreement would normally be required, the inflectional endings are simply absent:

16. svoj‑ega poceni avtomobil‑a
his‑gen.sg.m cheap‑uninfl car‑gen.sg.m
‘of his cheap car’

 

17. o t‑em poceni način‑u
about this‑loc.sg.m cheap‑uninfl method‑loc.sg.m
‘about this cheap method’

 

18. kupuje svoj‑o poceni letalsk‑o kart‑o
buys‑3sg.prs his‑acc.sg.f cheap‑uninfl airplane.adjacc.sg.f ticket‑acc.sg.f
‘he buys his cheap airplane ticket’

 

  • 27 Simplified from *cěnьn‑, the stem of the oblique forms of *cěnьnъ ‘*of a price’ ‘precious’/‘cheap (...)
  • 28 There are some more examples of this kind of development (prepositional phrase → uninflected adjec (...)

35This lexeme is also fully integrated into the morphology of standard Slovene in that it has a comparative and superlative (SSKJ2, s.v.), supplied by applying the productive comparative/superlative morphology to the doublet cẹ̑n ‘cheap’27 (cpv cen‑ȇjši, spv nàj‑cen‑ȇjši).28

  • 29 For a broader background and examples (in Croatian) of a somehwat different but related type, cf.  (...)

36A somewhat similar example of an adverbial element acquiring fully adjectival use, without any overt derivation or addition of inflectional endings, is colloquial B/C/M/S bèzvezē ‘irrelevant, meaningless, not making sense, silly’. It is a transparent univerbation of the prepositional phrase bèz vezē lit. ‘without connection’, i.e. the preposition bez followed by the noun vȅza ‘connection’ in the gen.sg; again, it can be used in any position in which ordinarily inflected adjectives occur:29

19. sv‑ih t‑ih bezveze zadatak‑a
all‑gen.pl.m that‑gen.pl.m silly‑uninfl task‑gen.pl.m
‘of all those silly tasks’

 

20. sa t‑om bezveze tehnik‑om
with this‑ins.sg.f silly‑uninfl technology‑ins.sg.f
‘with that silly technology’

 

  • 30 Example sourced from Google search by the author.

37This item is limited to the colloquial register, so that no dictionaries take a stance on its association with a derived comparative/superlative; still, one may find clear cases of the use of the derived synonym bèzvezan supplying its regularly formed comparative (bezvèznijī), e.g.:30

21. bezveze da bezveznij‑i biti ne može
silly‑uninfl that silly.adj‑cpv‑nom.sg.m be‑inf neg can‑3sg.prs
‘so silly that it cannot be sillier’

 

  • 31 Of course, such processes were not particularly common, predictable or uniform; other elements ari (...)

38We may note that the synonym bèzvezan (PSl. transposition *bezvęzьnъ) is derived with the suffix ‑an (< PSl. *‑ь‑nъ), i.e the same material that served to provide uninflected adjectives in *‑ь with inflected counterparts in *‑ь‑nъ from early on (see 1 above); i.e. B/C/M/S bèzvezē bèzvezan just like *svobodь ‘free‑uninfl’ → *svobodьnъ ‘free’.31

  • 32 (Examples sourced from Google searches by the author.) Agreement in Rom. adjectives is, of course, (...)

39Similar instances can, of course, also be cited for many other languages. One finds prepositional phrases and other adverbial expressions yielding fully integrated, uninflected adjectives e.g. in Romanian; cf. de‑adverbial adjectives like cumsecade ‘proper, honest’, a univerbation of the phrase cum se cade ‘as it is proper’ (with the 3sg.prs.refl of the verb a cădea ‘to fall, to be proper’). Again, an item like this may occur in positions where agreement would otherwise be necessary:32

22. datorită un‑or cumsecade oamen‑i
thanks to art.indefgen./dat.pl honest‑uninfl man‑pl.m
‘thanks to (some) honest people’

 

23. moart‑ea femei‑i cumsecade
death‑nom.sg.f.def woman‑gen./dat.sg.f.def honest‑uninfl
‘the death of the honest woman’

6. Towards a new approach

40Given the above parallels, it seems rather likely that a similar process underlies the adjectival use of the forms in *‑ь in early Sl. Just as adverbs originating in prepositional phrases came to function as indeclinable adjectives in the recent history of the modern Sl. languages (cf. Sln. pocẹ́ni, B/C/M/S bèzvezē discussed in the previous section), the PSl. adverbs in *‑ь apparently developed such use too.

41The choice of prepositional phrases for the typological parallels mentioned before is not coincidental: the Sl. forms in *‑ь (both in their adverbial and adjectival guise) are very closely connected with prepositional phrases. As already mentioned above, adverbial univerbations of prepositional phrases substitute *‑ь in the second element (on the phenomenon see e.g. Stang 1973): cf. OCS slědъ ‘trace, footprint’ → poslědь ‘afterwards’ (doublet of prepositional phrase po slědu, lit. ‘following the footprint’), OCS kupa ‘heap, multitude’ → vъkupь ‘in a multitude’ (doublet of prepositional phrase vъ kupě ‘in a multitude’). Besides, in certain rarer cases, a derived form in *‑ь appears to occur in lieu of the base noun in a prepositional phrase; cf. ORu. usta ‘mouth’ (pl.tant.n) → na ustь rěky ‘at the mouth of the river’.

  • 33 On this famous phenomenon and different takes on its prehistory in both Sl. and Balt., cf. Koch 19 (...)

42Another interesting point may be raised regarding the distributional properties of the forms in *‑ь. One of the hallmarks of the Sl. adjective is the occurrence of so‑called “long” or “definite” forms (i.e. the coupling with a pronominal element *) beside plain ‘short’ or ‘indefinite’ forms.33 Note that long/definite forms with *are in fact attested both for uninflected adjectives in *‑ь and for prepositional phrases; cf. OCS:

24. svobodьi
/svobodь jь/
free.uninfl def‑nom.sg.m
‘the free one’

 

25. nabožijǫi člověkъ (cf. Koch 1992, p. 64)
/na božьjǫ člověkъ/
in godly‑acc.sg.f def‑nom.sg.m man‑nom.sg.m
‘the pious man’ (lit. ‘the in‑a‑godly‑[manner] man’)

 

  • 34 Needless to say, the division into parts of speech is often problematic and fluid in itself, even (...)
  • 35 Such as e.g. in Middle (and Modern) Germ., largely due to diachronic phonological reductions in fi (...)

43As concerns Vaillant’s argument that uninflected adjectives in *‑ь cannot originate in the homophonous adverbs in *‑ь because such adjectival use “n’apparaît pas avec des adverbes d’autres types”, it seems that this reservation is mostly an artifact of the extremely broad scope of the concept “adverb”. As is well‑known, adverbs generally constitute perhaps the least internally consistent part‑of‑speech category.34 Drawing the boundary between the adverb and other parts of speech is problematic not only in those languages where the formal means for signaling the distinction are reduced due to general morphological impoverishment;35 the problem is much broader and essentially cross‑linguistic. Cf. e.g.:

Adverb is a “catch‑all” category. Any word with semantic content (i.e., other than grammatical particles) that is not clearly a noun, a verb, or an adjective is often put into the class of adverb […]. [Adverbs] are typically the most unrestricted category in terms of their position in clauses. (Payne 1997, p. 69)

Was also ist ein Adverb? Wir können die Frage nicht beantworten, weil wir den Verdacht hegen, daß es sich um eine weitgehend undefinierbare Wortklasse handelt. (Schwarz 1982, p. 64)

  • 36 We may note that under some approaches, the distinction between adverbs and adjectives is in fact (...)

44The present study is, of course, no place for a typological investigation into what constitutes an adverb.36 Note, however, that in the languages from which the above parallels were drawn, it is also not the case that any adverbial expression (let alone any closed‑class adverb) may come to be used adjectivally. Presumably, such behavior can be expected from words denoting more “prototypically adjectival” qualities (cf. e.g. Dixon 1997; 2004). Be that as it may, the fact that one class of items describable as “adverbs” does not display a given kind of patterning in no way implies that another class of such items cannot do it either; on the contrary, such differences in distribution are only to be expected in view of the internal diversity of the category. Hence, Vaillant’s “n’apparaît pas avec des adverbes d’autres types” cannot be considered a crippling argument.

45As regards the trajectory via which adverbs in *‑ь may have come to be used adjectivally, we may suspect that predicative position was the pivotal point; from there, the forms would have been extended to attributive use (on this type of development cf. Hoffmann 1952). The same process presumably underlies the modern parallels mentioned in 6 above.

7. Conclusions and reconstructed scenario

46On the basis of the discussion presented in the preceding sections, we are in a position to devise an account of the Sl. uninflected adjectives in *‑ь that does not require a surprising loss of inflection of any kind. The occurrence of invariable adjectival forms in *‑ь in Sl. can be explained assuming the following chain of rather unremarkable diachronic developments:

  1. the nom./acc.sg.n of the PIE compounding type ákṣiti‑/inermis/ἄναλκις was extracted from the paradigm as a productive adverbial form in Pre‑PSl. (thus simply increasing the productivity of a PIE‑age model, cf. e.g. Gr. ἀμισθί ‘for free’, adv. of ἄμισθος ‘unpaid’, from ἀ‑ ‘un‑’ and μισθός ‘payment’), giving rise to the productive PSl. adverbs in *‑ь; these often served as univerbations or doublets of prepositional phrases;
  2. at an indeterminable Post‑PBSl. but Pre‑PSl. stage, i‑stem adjectives themselves were lost as a category (save, perhaps, for *gorьkъ ‘bitter’ and *tęžьkъ ‘heavy’, if from a genuine *‑i‑ko‑);
  3. at some later point, adverbs in *‑ь started being used in an adjective‑like fashion (full capability of attributive, adnominal use); this was entirely parallel to the later integration of other adverb‑ or prepositional phrase‑like structures into the system of adjectives in the individual Sl. languages (type Sln. pocẹ́ni, B/C/M/S bèzvezē) as well as of the earliest Sl. itself (type OCS nabožijǫi člověkъ).
  • 37 A similar scenario can perhaps be read into certain older works, e.g. Brugmann 1906, p. 112: “‑i‑: (...)

47An important consequence is that – pace the standard handbooks – Sl. does not attest PIE i‑stem adjectives as such at all. The commonly found wording “PIE i‑stem adjectives became uninflected in Slavic”, or similar, is ambiguous and misleading; by default, it points to a diachronic innovation wholly unexpected in early Sl., i.e. a loss of inflection in a single class of adjectives. Thus, the statement should be replaced by something along the lines of: “an adverbial form derived from the nom./acc.sg.n of PIE i‑stem adjectives developed adjectival use in Slavic”.37

  • 38 See 1 above. As regards adverbs in *‑i, one wonders whether forms of such type did not also contri (...)

48Finally, we may note that the scenario has practically no bearing on Baltic, where i‑stem adjectives were indeed retained as an actual inflectional class.38

Bibliographie

Abbreviations

Beekes EDG: R.S.P. Beekes, Etymological dictionary of Greek, vol. I‑II, Leiden/Boston, Brill, 2010.

ESJS: E. Havlová, Etymologický slovník jazyka staroslověnského, vol. I‑XIX, Prague, Academia, 1989‑2018.

Schuster‑Šewc HEW: H. Schuster‑Šewc, Historisch‑etymologisches Wörterbuch der ober‑ und niedersorbischen Sprache, vol. I‑V, Bautzen, VEB Domowina, 1978‑1994.

SJS: J. Kurz, Slovník jazyka staroslověnského. Lexicon linguae palaeoslovenicae, vol. I‑IV, Prague, Nakl. Československé akademie věd, 1958‑1997.

Snoj SES3: M. Snoj, Slovenski etimološki slovar, Ljubljana, Založba ZRC, 2016 (3rd ed.), https://fran.si/193/marko-snoj-slovenski-etimoloski-slovar (accessed 01/09/2019).

SSKJ2: Slovar slovenskega knjižnega jezika, Ljubljana, Cankarjeva založba, 2014 (2nd ed.), https://fran.si/133/sskj2-slovar-slovenskega-knjiznega-jezika-2 (accessed 01/09/2019).

Vasmer ÈSRJa : M. Fasmer [Vasmer], Ètimologičeskij slovar’ russkogo jazyka. Perevod s nemeckogo i dopolnenija člena‑korrespondenta AN SSSR O. N. Trubačëva, vol. I‑IV, Moscow, Progress, 1986‑1987.

References

Aitzetmüller 1991: R. Aitzetmüller, Altbulgarische Grammatik als Einführung in die slavische Sprachwissenschaft, Freiburg i.Br., U.W. Weiher, 1991 (2nd ed.).

Arumaa 1985: P. Arumaa, Urslavische Grammatik. Einführung in das vergleichende Studium der slavischen Sprachen, vol. III, Formenlehre, Heidelberg, Winter, 1985.

Backhouse 2004: A. Backhouse, “Inflected and uninflected adjectives in Japanese”, in R.M.W. Dixon, A.Y. Aikhenvald (ed.), Adjective classes. A cross‑linguistic typology, Oxford/New York, Oxford University Press, 2004, pp. 50‑73.

Balles 2006: I. Balles, Die altindische Cvi‑Konstruktion. Form ‑ Funktion ‑ Ursprung, Bremen, Hempen, 2006.

Balles 2009: I. Balles, “Zu den i‑stämmigen Adjektiven des Lateinischen”, in R. Lühr, S. Ziegler (ed.), Protolanguage and Prehistory. Akten der XII. Fachtagung der Indogermanischen Gesellschaft vom 11. bis 15. Oktober 2004 in Krakau, Wiesbaden, Reichert, 2009, pp. 1‑26.

Barteld 2015: F. Barteld, “On the distinction between adverbs and adjectives in Middle High German”, in K. Pittner, F. Barteld, D. Elsner (ed.), Adverbs. Functional and diachronic aspects, Amsterdam/Philadelphia, John Benjamins, 2015, pp. 157‑178.

Brugmann 1906: K. Brugmann, Grundriß der vergleichenden Grammatik der indogermanischen Sprachen, vol. II, Lehre von Wortformen und ihrem Gebrauch, part 1, Allgemeines. Zusammensetzung (Komposita). Nominalstämme, Strasbourg, Trübner, 1906.

Diels 1963: P. Diels, Altkirchenslavische Grammatik, Heidelberg, Winter, 1963 (2nd ed.).

Dixon 1977: R.M.W. Dixon, “Where have all the adjectives gone?”, Studies in Language 1, 1977, pp. 19‑80.

Dixon 2004: R.M.W. Dixon, “Adjective classes in typological perspective”, in R.M.W. Dixon, A.Y. Aikhenvald (ed.), Adjective classes. A cross‑linguistic typology, Oxford/New York, Oxford University Press, 2004, pp. 1‑49.

Doleschal 2015: U. Doleschal, “Završena velika Konzum nagradna igra″ – on the status of premodifying nouns in Croatian”, Fluminensia 27/2, 2015, pp. 191‑202.

Dunkel 2014: G. Dunkel, Lexikon der indogermanischen Partikeln und Pronominalstämme, vol. II, Lexikon, Heidelberg, Winter, 2014.

Duridanov et al. 1991: I. Duridanov, E. Dogramadžieva, A.C. Minčeva, I. Bujukliev (ed.), Gramatika na starobălgarskija ezik. Fonetika, morfologija, sintaksis, Sofia, Izdatelstvo na Bălgarskata akademija na naukite, 1991.

Forssman 2003: B. Forssman, Das baltische Adverb. Morphosemantik und Diachronie, Heidelberg, Winter, 2003.

Greenberg 2006: M. Greenberg, A short reference grammar of standard Slovene, Lawrence, SEELRC, 2006, http://hdl.handle.net/1808/5469 and https://fran.si/slovnice-in-pravopisi/51/20082006-greenberg (accessed 01/09/2018).

Grestenberger 2014: L. Grestenberger, “Zur Funktion des Nominalsuffixes *‑i‑ im Vedischen und Urindogermanischen”, in N. Oettinger, T. Steer (ed.), Das Nomen im Indogermanischen. Morphologie, Substantiv versus Adjektiv, Kollektivum. Akten der Arbeitstagung der Indogermanischen Gesellschaft vom 14. bis 16. September 2011 in Erlangen, Wiesbaden, Reichert, 2014, pp. 88‑102.

Grestenberger 2017: L. Grestenberger, “On ‘i‑substantivization’ in Vedic compounds”, in B.S.S. Hansen, A. Hyllested, A.R. Jørgensen, G. Kroonen, J.H. Larsson, B.N. Whitehead, T. Olander, T.M. Søborg (ed.), Usque ad Radices. Indo‑European studies in honour of Birgit Anette Olsen, Copenhagen, Museum Tusculanum Press, 2017, pp. 193‑206.

Grošelj 2007: R. Grošelj, “Miklošičev prispevek k skladenjski obravnavi starocerkvenoslovanskega nepredložnega mestnika”, Slavistična revija 55, 2007, pp. 105‑131.

Hengeveld 2005: K. Hengeveld, “Parts of speech”, in M. Anstey, J.L. Mackenzie (ed.), Crucial readings in functional grammar, Berlin/Boston, De Gruyter, 2005, pp. 79‑106.

Hoffmann 1952: K. Hoffmann, “Zum prädikativen Adverb”, Münchener Studien zur Sprachwissenschaft 1, 1952, pp. 42‑53 (reprint in K. Hoffmann, J. Narten (ed.), Aufsätze zur Indoiranistik, vol. II, Wiesbaden, Reichert, 1976, pp. 339‑349).

Höfler 2019: S. Höfler, “The two types of ‘secondary’ *‑(e‑)h2 stems in PIE”, paper presented at: Ljubilej – Ljubljäum – Ljubilee – Ljubljanniversaire. IG / SIES / SÉIE Arbeitstagung, Ljubljana, 4‑7 June 2019, (unpublished).

Jaros 2016: I. Jaros, “Przymiotnikowe derywaty o znaczeniu niepełnej cechy i ich prefigowane synonimy w gwarach Polski centralnej”, Rozprawy Komisji Językowej ŁTN / Dissertations of Language Committee of Lodz Learned Society 63, 2016, pp. 17‑34.

Jelitte 1961: H. Jelitte, Studien zum Adverbium und zur adverbialen Bestimmung im Altkirchenslavischen, Meisenheim am Glan, Hain, 1961.

Klingenschmitt 1992: G. Klingenschmitt, “Die lateinische Nominalflexion”, in O. Panagl, T. Krisch (ed.), Latein und Indogermanisch. Akten des Kolloquiums der Indogermanischen Gesellschaft, Salzburg, 23.‑26. September 1986, Innsbruck, Institut für Sprachwissenschaft, 1992, pp. 89‑135.

Koch 1992: C. Koch, “Zur Vorgeschichte des relativen Attributivkonnexes im Baltischen und Slavischen”, in B. Barschel, M. Kozianka, K. Weber (ed.), Indogermanisch, Slawisch und Baltisch. Materialien des vom 21.‑22. September 1989 in Jena in Zusammenarbeit mit der Indogermanischen Gesellschaft durchgeführten Kolloquiums, Munich, Peter Lang, 1992, pp. 45‑88.

Koneveckij 1977: A.K. Koneveckij, “O russkix narečijax, vosxodjaščix k imenitel’nomu padežu imeni”, Kalbotyra 28/4, 1977, pp. 51‑60.

Krupska‑Perek 1988: A. Krupska‑Perek, “Ludowe cechy słowotwórstwa przymiotników w gwarach centralnej Polski”, Rozprawy Komisji Językowej ŁTN / Dissertations of Language Committee of Lodz Learned Society 34, 1988, pp. 97‑103.

Le Feuvre 2010: C. Le Feuvre, “Suffixation et composition: composés et dérivés de la racine *Heu̯‑ ‘voir’ dans les langues indo‑européennes”, Bulletin de la Société de Linguistique de Paris 105, 2010, pp. 125‑144.

Malinowski 2016: P. Malinowski, “O zaimku ów”, Obcy język polski, 2016, https://obcyjezykpolski.pl/o-zaimku-ow/ (accessed 01/09/2018).

Markowski 2004: A. Markowski (ed.), Wielki słownik poprawnej polszczyzny, Warsaw, Wydawnictwo Naukowe PWN, 2004.

Meier‑Brügger 2010: M. Meier‑Brügger, Indogermanische Sprachwissenschaft, Berlin/New York, De Gruyter, 2010 (9th ed.).

Meillet 1905: A. Meillet, Études sur l’étymologie et le vocabulaire du vieux slave, t. II, Paris, Bouillon, 1905.

Meillet 1934: A. Meillet, Le slave commun, Paris, E. Champion, 1934 (2nd ed.).

Miklošič 1868‑1874: F. Miklošič, Vergleichende Grammatik der slavischen Sprachen, vol. IV, Syntax, Vienna, W. Braumüller, 1868‑1874.

Odijk 1992: J. Odijk, “Uninflected adjectives in Dutch”, in R. Bok‑Bennema, R. van Hout (ed.), Linguistics in the Netherlands, Amsterdam, John Benjamins, 1992, pp. 197‑208.

Olander 2015: T. Olander, Proto‑Slavic inflectional morphology. A comparative handbook, Leiden, Brill, 2015.

Orzechowska 1963: A. Orzechowska, “Kilka uwag o przymiotnikach nieodmiennych i o pokrewnych zjawiskach w języku słoweńskim”, International Journal of Slavic Linguistics and Poetics 7, 1963, pp. 18‑51.

Payne 1997: T. Payne, Describing morphosyntax. A guide for field linguists, Cambridge/New York, Cambridge University Press, 1997.

Perzowa 1969: H. Perzowa, Stopniowanie przymiotników polskich z przedrostkiem przy‑, Wrocław, Ossolineum, 1969.

Petit 1999: D. Petit, “Introduction au lituanien”, Lalies 19, 1999, pp. 7‑135.

Petit 2009: D. Petit, “La préhistoire des adjectifs déterminés du baltique et du slave”, Bulletin de la Société de Linguistique de Paris 104, 2009, pp. 311‑360.

Petit 2011: D. Petit, “On the etymology of the Latvian comparative vaĩrs, vaĩrâk”, Studia Etymologica Cracoviensia 16, 2011, pp. 103‑118.

Rau 2009: J. Rau, Indo‑European nominal morphology. The decads and the Caland system, Innsbruck, Institut für Sprachen und Literaturen der Universität Innsbruck, 2009.

Rau 2014: J. Rau, “The history of the Indo‑European primary comparative”, in N. Oettinger, T. Steer (ed.), Das Nomen im Indogermanischen. Morphologie, Substantiv versus Adjektiv, Kollektivum. Akten der Arbeitstagung der Indogermanischen Gesellschaft vom 14. bis 16. September 2011 in Erlangen, Wiesbaden, Reichert, 2014, pp. 327‑241.

Reindl 2008: D. Reindl, Language contact: German and Slovenian, Bochum, Brockmeyer, 2008.

Schwarz 1982: C. Schwarz, “Was ist ein Adverb?”, Linguistiche Berichte 81, 1982, pp. 61‑65.

Sims‑Williams, Enger 2021: H. Sims‑Williams, H.‑O. Enger, “The loss of inflection as grammar complication. Evidence from Mainland Scandinavian”, Diachronica 38, 2021, pp. 111‑150.

Sobotka 2006: P. Sobotka, “ów”, Poradnia językowa PWN, 2006, https://sjp.pwn.pl/poradnia/haslo/;7581 (accessed 01/09/2019).

Sommer 2016: F. Sommer, “The historical morphology of definiteness in Baltic”, Indo‑European Linguistics 6, 2016, pp. 152‑200.

Stang 1939: C. Stang, “Slavische indeklinable Adjektiva auf ‑ь”, Norsk tidsskrift for sprogvidenskap 11, 1939, pp. 99‑103.

Stang 1966: C. Stang, Vergleichende Grammatik der baltischen Sprachen, Oslo, Universitetsforlaget, 1966.

Stang 1973: C. Stang, “Sur la mutation en ‑i‑ dans la formation des noms en slave et en baltique”, Norwegian Journal of Linguistics/Norsk Tidsskrift for Sprogvidenskap 27, 1973, pp. 77‑84.

Stieber 1979: Z. Stieber, Zarys gramatyki porównawczej języków słowiańskich, Warsaw, Państwowe Wydawnictwo Naukowe, 1979.

Vaillant 1948: A. Vaillant, Manuel du vieux slave, t. I, Grammaire, Paris, Institut d’Études slaves, 1948.

Vaillant 1956: A. Vaillant, “Etymologica (Vieux slave spyti – et tunje, pol. tani)”, Die Welt der Slaven 1/2, 1956, pp. 140‑142.

Vaillant 1958: A. Vaillant, Grammaire comparée des langues slaves, t. II, Morphologie, part 1, Flexion nominale, part 2, Flexion pronominale, Paris, IAC, 1958.

Villanueva‑Svensson 2018: M. Villanueva‑Svensson, “The conditioning of the Balto‑Slavic i‑apocope”, Münchener Studien zur Sprachwissenschaft 71/2, 2018, pp. 181‑200.

Villanueva‑Svensson 2019: M. Villanueva‑Svensson, “The infinitive in Baltic and Balto‑Slavic”, Indo‑European Linguistics 7/1, 2019, pp. 194‑221.

Vogel 1996: P. Vogel, Wortarten und Wortartenwechsel. Zur Konversion und verwandten Eigenschaften im Deutschen und in anderen Sprachen, Berlin/New York, De Gruyter, 1996.

Vondrák 1924: V. Vondrák, Vergleichende slavische Grammatik, vol. I, Lautlehre und Stammbildungslehre, Göttingen, Vandenhoeck & Ruprecht, 1924 (2nd ed.).

Van Wijk 1931: N. van Wijk, Geschichte der altkirchenslavischen Sprache, vol. I, Laut‑ und Formenlehre, Berlin, De Gruyter, 1931.

Willis 2008: D. Willis, “Degrammaticalization, exaptation and loss of inflection: evidence from Slavonic”, Cambridge, 2008, https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.123.6796&rep=rep1&type=pdf (accessed 18/02/2021).

Xaburgaev 1986: G.A. Xaburgaev, Staroslavjanskij jazyk, Moscow, Prosveštenie, 1986.

Annexes

Abbreviations and symbols

acc

accusative

adj

adjective

adv

adverb

Alb.

Albanian

Balt.

Baltic

B/C/M/S

Bosnian/Croatian/Montenegrin/Serbian

Bg.

Bulgarian

cpv

comparative

Cz.

Czech

dat

dative

def

definite

du

dual

Eng.

English

f

feminine

Fr.

French

fut

future

gen

genitive

Germ.

German

Goth.

Gothic

Gr.

Ancient Greek

Hitt.

Hittite

IE

Indo‑European

indef

indefinite

inf

infinitive

ins

instrumental

Lat.

Latin

lit.

literally

Lith.

Lithuanian

loc

locative

LSorb.

Lower Sorbian

m

masculine

n

neuter

neg

negation

nom

nominative

O

Old

OCS

Old Church Slavic

OE

Old English

OIr.

Old Irish

OPr.

Old Prussian

PIE

Proto‑Indo‑European

PIran.

Proto‑Iranian

pl

plural

Pol.

Polish

prep

preposition

pron

pronoun

prs

present

refl

reflexive

rel

relative

Rom.

Romanian

Ru.

Russian

sg

singular

Sl.

Slavic

Sln.

Slovene

spv

superlative

subst

substantive

tant

(plurale) tantum

Ukr.

Ukrainian

uninfl

uninflected

USorb.

Upper Sorbian

Ved.

Vedic Sanskrit

borrowed/derived from

borrowed/derived into

<

developed phonologically from

>

developed phonologically into

<<

developed from (with morphological change)

>>

developed into (with morphological change)

*

reconstructed form

**

ungrammatical form

#

substandard form

Notes

1 I would like to thank the participants of the Rouen conference for their helpful remarks and suggestions. Aspects of this research have also been presented in guest talks in Cologne (May 2018) and Vilnius (December 2019), as well as in a contribution to the conference “Przeszłość w języku zamknięta. In memoriam Andreae Bańkowski II” (Częstochowa, 22‑23 October 2018); participants of all these events likewise provided valuable input. I alone am responsible for all opinions expressed in the present article.

2 Note that this fully overlapping distribution makes the Sl. phenomenon quite distinct from seemingly similar dichotomies known from certain other languages, famously e.g. the ‘inflected’/‘uninflected’ adjective classes in Japanese (cf. Backhouse 2004), or somewhat less famously e.g. in Dutch (cf. Odijk 1992).

3 For example, traces are extant in certain varieties of Pol. until the 18th century (cf. Jaros 2016, p. 19; Krupska‑Perek 1988, p. 101; Perzowa 1969, p. 60): e.g. obstarz ‘rather old’ (as though < *‑starь; cf. Pol. stary ‘old’ < PSl. *starъ), obdłuż ‘rather long’ (as though < *‑dьlžь < *‑dьlgь; cf. Pol. długi ‘long’ < PSl. *dьlgъ), etc., attested in uninflected form in attributive position under various number, gender and case environments. In the dialects in question, these forms later got replaced by derivatives in ‑ni: obstarni, obdłużni etc. (regular and fully inflected).

4 However, as is well‑known, the status of the category in PIE is somewhat elusive, since direct equations of i‑stem adjectives across the IE family are hard to come by. While it seems certain that some types of derived i‑stem adjective formations do go back to the proto‑language, the question whether *‑i‑ could also function as a primary adjectival suffix in PIE (like *‑u‑, *‑ro‑ and other elements associated with the Caland System) remains controversial. Recently on the status of i‑stem adjectives in PIE see e.g. (all with further literature): Balles 2009, pp. 19‑20 and passim; Rau 2009, pp. 72 and 132; 2014, p. 339; Meier‑Brügger 2010, p. 354; Grestenberger 2014, pp. 94‑95; most lately Höfler 2019, p. 12. See also further below.

5 If a substantivized neuter of *dʰowgʰ‑i‑ – probably the morphological type of Gr. τρόφις built to the root *dʰewgʰ‑ ‘produce, obtain’; thus e.g. Stang 1966, p. 320; Arumaa 1985, p. 55; Petit 2011, p. 116. Alternative (e.g. verbal) interpretations of this word exist. Even if the entirety of the Balt. evidence for i‑stem adjectives is dismissed, however, the theory presented here requires no substantial modifications.

6 Two PSl. adjectives – *tęžьkъ ‘heavy’ and *gorьkъ ‘burning, bitter’ – have been suspected of continuing earlier i‑stem forms as well, but remodeled in a fashion parallel to that observed in the u‑stems, i.e. with the addition of the suffix *‑kъ (< *‑ko‑) in the positive form (as though *t(ʰ)n̥gʰ‑i‑ko‑, *gʷor‑i‑ko‑). According to a different view, however, the forms in *‑ь‑kъ are late and replace original *‑ъ‑kъ, i.e. u‑stem *tęgъkъ and *gorъkъ (*‑u‑ko‑). The main argument for the u‑stem reconstruction comes from the fact that *tęgъkъ is indirectly attested in the derivative OCS *otęgъčati ‘become heavy’ (3pl.prs otęgъčajǫtъ); cf. also the directly superimposable Lith. tingùs ‘lazy’. Under this theory, the change of *‑ъ‑kъ to *‑ь‑kъ has been ascribed to the influence of comparative forms (*tęž‑ьš‑ ‘heavier’, *gor’‑ьš‑ ‘*more bitter > worse’), formed deradically with the suffix *‑jьš‑ (<< *‑yos‑, *‑is‑) and displaying palatalization of the stem. The palatalized stem would have been analogically transferred to the positive, with concomitant regular fronting of the following *‑ъ‑ to *‑ь‑ in the positive suffix (not inducing the 3rd palatalization of the velar, however). Whether *tęžьkъ and *gorьkъ are real morphological reflexes of i‑stem adjectives or late modifications of u‑stem forms in *‑ъkъ is of no great importance for the present purposes, i.e. the explanation of the uninflected forms. The question does, however, have some bearing on the precise description of the fate of i‑stem adjectives in Sl.; see 7 below.

7 Cf. PSl. substantival i‑stem nom.pl.m *‑ьje, as e.g. in OCS tatie ‘thieves’ (either directly < PIE *‑ey‑es or from the stacking of PSl. *‑e [< PIE *‑es] onto PSl. *‑i‑ [< PIE *‑eye‑]; the decision depends on the assumed phonological and morphological history. Discussion in Olander 2015, pp. 224‑226).

8 Cf. PSl. substantival i‑stem dat.sg.m *‑i, as e.g. in OCS pǫti ‘way’ (either < PIE *‑ey‑ey or haplologized *‑ey; the decision depends on the assumed phonological and morphological history. Discussion in Olander 2015, pp. 145‑147; Klingenschmitt 1992, pp. 105‑108; most recently Villanueva‑Svensson 2019).

9 Cf. PSl. substantival i‑stem nom.du.f *‑i, as e.g. in *kosti ‘bones’ (< PIE *‑i‑h1; discussion in Olander 2015, pp. 191‑192).

10 See, however, fn. 6 on the two adjectives (*tęžьkъ ‘heavy’ and *gorьkъ ‘burning, bitter’) which have been suspected of adhering to this tendency (*‑i‑ remodeled to *‑i‑ko‑).

11 Whether such an innovation would have indeed ‘repaired’ the grammar in this context is far from obvious; cf. Sims-Williams, Enger 2021 for thought‑provoking discussion on whether indeclinability universally amounts to ‘simplicity’.

12 On definite/long forms – including in the context of the uninflected adjectives in *‑ь – see also 6 below.

13 On the IE etymology cf. e.g. ESJS, vol. X, p. 613; Dunkel 2014, p. 111. PSl. *ovъ is usually treated as continuing a (Post‑)PIE demonstrative *aw‑o‑, *h2ew‑o‑ or similar, probably derived from an underlying adverbial/presentative element. The presumed cognates include PIran. *au̯a‑ ‘that, jener’, Gr. αὐτός ‘that, he, himself, ipse’ (*aw‑to‑, though conceivably heterogeneous in view of the polyfunctional nature), Alb. ai ‘that, he’ (*aw‑so), etc.

14 Constructed by the author as standard‑usage counterparts of ‎(10)‑‎(12) below.

15 In the inflectional morphology of Pol. (and most of modern Sl.), usually acc.sg.m = nom.sg.m in non‑animate nominal phrases, but acc.sg.m = gen.sg.m in animate nominal phrases.

16 Examples sourced from Google searches by the author; further examples available (with discussion) in the references following in the main text.

17 As discussed at some length in the references provided above.

18 In older Sln., the pronoun also occurs in the variant kir; it also partly continues the PSl. relative *jьže, with substitution of *k‑ for *j‑ in the stem and the regular Western South Slavic rhotacism of *ž to r.

19 Examples sourced from Google searches by the author.

20 Except for the derivation of by‑forms in *‑ь‑nъ (< *‑i‑no‑) etc.; see in 1 above.

21 For some further Sl. examples of this type cf. Willis 2008.

22 This example, though widely used to illustrate the type, is somewhat problematic in view of the evidence for the root noun ἀλκ‑ (cf. Beekes EDG, vol. I, p. 70). The derivational model itself, however, is non‑controversial.

23 Thus, e.g. the adjectives *pravъ ‘straight’ and *prokъ ‘remaining’ referred to above are compounds from the etymological point of view, though no doubt fully opaque by the PSl. period. On the etymology and compounding model cf. Le Feuvre 2010.

24 The alternative analysis by Meillet (1905, p. 266), according to which the Sl. forms in *‑ь are historically yo‑stems (but with nom.sg *‑i‑s and acc.sg *‑i‑m, recalling the surface reflexes of yo‑stem nom.sg.m and acc.sg.m in Lith. or Goth.), is purely of historical interest today, as judged already by Stang 1939, p. 100. (Note that in 1934, p. 471, Meillet does not mention this theory anymore, calling the forms in *‑ь a “type […] dont l’origine est tout à fait obscure”). Other analyses, e.g. Koneveckij 1977, are also difficult to accept. Finally, it should be noted that the type is certainly not entirely homogenous; some classes of PSl. adverbs in *‑ь evidently come from other sources, e.g. the consonant stem loc.sg *‑i and other forms (see Villanueva‑Svensson 2018).

25 Cf. fn. 24 towards the end.

26 For the implication of a good/low/favorable price in the foundational phrase, cf. the semantic development of Eng. cheap, which likewise derives from the concept of ‘purchase’ alone (OE cēap ‘purchase, trade, value, price, etc.’, ultimately ← Lat. caupō ‘merchant’), with the favorable aspect implied but not overtly expressed.

27 Simplified from *cěnьn‑, the stem of the oblique forms of *cěnьnъ ‘*of a price’ > ‘precious’/‘cheap’ (Bg. cènen, Ru. cénnyj, Cz. cenný, etc.).

28 There are some more examples of this kind of development (prepositional phrase → uninflected adjective) in Sln. Most other uninflected adjectives in Sln. are borrowings, however – be it recent (e.g. from Eng. or Fr.) or historical (mostly from Germ.). For discussion and references, cf. Reindl 2008, pp. 78‑80 (who links the entrenchment of the type to prolonged contact with Germ.); cf. also Orzechowska 1963. Indeclinable adjectives resulting from recent borrowings are of course a commonplace occurrence in all Sl. languages.

29 For a broader background and examples (in Croatian) of a somehwat different but related type, cf. Doleschal 2015 (similar developments can also be found in Slovene, cf. Greenberg 2006, p. 42).

30 Example sourced from Google search by the author.

31 Of course, such processes were not particularly common, predictable or uniform; other elements arising in a similar way in Slavic did receive productive adjectival morphology, thus becoming regular, inflectable adjectives. Cf. e.g. Vaillant’s (1956, p. 142) account of the adjective *tun’ь var. *tan’ь ‘cheap’ (USorb. tuni, LSorb. tuny ‘cheap’; adv OCS tunje, Ru. túne ‘in vain’ etc.; Pol. tani ‘cheap’), argued to derive from an expression *tu na (var. *ta na) ‘là, tiens!’ or similar (“[c]ette expression invitait la personne à laquelle on s’adressait [tu « là près de toi », 2e personne] à prendre ce qu’on lui offrait, et consacrait le cadeau”). The adjective would have been derived with the suffix *‑jь. For different explanations of *tunь/*tanь see e.g. Snoj SES3, s.v. and Schuster‑Šewc HEW, s.v.; Vasmer ÈSRJa, s.v. regards the item as unexplained.

32 (Examples sourced from Google searches by the author.) Agreement in Rom. adjectives is, of course, far less developed than in any medieval Sl. language; still, most adjectives of non‑adverbial origin do exhibit basic agreement for number, gender and case.

33 On this famous phenomenon and different takes on its prehistory in both Sl. and Balt., cf. Koch 1992; Petit 2009; Sommer 2016.

34 Needless to say, the division into parts of speech is often problematic and fluid in itself, even concerning the more established categories; see e.g. Vogel 1996.

35 Such as e.g. in Middle (and Modern) Germ., largely due to diachronic phonological reductions in final syllables; see e.g. Barteld 2015.

36 We may note that under some approaches, the distinction between adverbs and adjectives is in fact reduced to distributional properties, with adjectives modifying nominals and adverbs modifying other types of heads (cf. Hengeveld 2005, p. 87). For more discussion on the cross‑linguistic and typological status of adverbs, cf. the references adduced in the present section, with copious further literature.

37 A similar scenario can perhaps be read into certain older works, e.g. Brugmann 1906, p. 112: “‑i‑: ai. dhūmá‑gandhi‑ṣ̌ ‘nach Rauch riechend’ (gandhá‑s) […] gr. ἄν‑αλκις ‘ohne Kraft’ (ἀλκή) ‘schwach’, lat. com‑mūnis zu lit. maĩnas ‘Tausch’ […]. Aus dem Slav. hierher die kompositionellen Formen auf ь (Nom.‑Akk. Sg. N.), die teils indeklinabel als Adjektiva, teils als Adverbia gebraucht sind, z. B. isplъnь ‘voll’, prěprostь ‘einfach’ […]”. It appears that Brugmann’s wording is at least compatible with the trajectory proposed in the present paper. He does not specify clearly which usage he considers older (“teils indeklinabel als Adjektiva, teils als Adverbia gebraucht”), but the reference to the “Nom.‑Akk. Sg. N.” seems to imply that the adverbial function can be considered original.

38 See 1 above. As regards adverbs in *‑i, one wonders whether forms of such type did not also contribute – via apocope of *‑i – to the endingless adverbial type of Balt. (Lith. šveñtadien ‘on Sunday’ ← šveñtadienis ‘Sunday’, vãkar ‘yesterday’ ← vãkaras ‘evening’ etc.) – a somewhat unclear and certainly heterogeneous class (see Forssman 2003).

Auteur

University of Lodz, Department of Slavic Philology, Poland

CC-BY-NC-ND-4.0

Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search