Between lexicon and grammar finding the way Vladimir

Between lexicon and grammar: finding the way Vladimir Plungian (Moscow)

Lexical vs. grammatical • Studying lexicon is studying “words”: typically, full-fledged autonomous units (or their combinations) – “minimal free forms”, in traditional structuralist parlance • Studying grammar is studying something “less” than words: typically, “bound” forms (= affixes)

Lexical vs. grammatical • In fact, we have two largely independent properties here: – “bound” vs. “free” – “word” vs. “word part” • Word parts are usually “bound” (though marginal counterexamples do exist) – Occurrences of contra- and pro-hedonic orientations were not mutually exclusive <…> [The Oxford Handbook of Emotion, 2014, 69 -70] • However, full-fledged words (not clitics!) can be “bound” as well

“Bound words” • Require obligatory modifiers and/or complements (= cannot or at least tend not to occur as a unique member of a syntactic unit) • Free words: moon, stop, mine, you, … • Bound words: because, yet, my, requires, not, … • “Weakly autonomous” units

Lexical vs. grammatical • Grammatical meanings can be expressed by: – non-segmental markers (phonemic and tonal alternations, reduplication, conversion) – affixation of various kinds – clitics – bound words (“auxiliaries”, “ancillary words”) • Most likely, not by free words • Still, the intersection between what is accessible both to lexical and grammatical domains is significant

Lexical affixes and clitics? • Andi [Majsak 2017, ex. 15] hege-l. u-w mič’i=ʁib. š-di <. . . > hit’on. . . [DEM-PL-GEN(M) little]=ATR. CONTR-ERG say. AOR ‘The younger of them said. . . ’ (Luke 15: 12) ʁib – ‘exactly, just’: an “illicit” inclusion of a lexical discourse unit into a well-formed nominal word-form

Lexical affixes and clitics? • Uzbek [Guzev 2004] šu šahar-dan-lig-im [this town-ABL]-NMLZ-POSS 1: SG ‘(the fact) that I am from this town’ » ‘my from-this-town-ness’ šahar-dan-lig-im : a possessive marker applies to what is better thought of as a free syntactic unit NB! derivation post inflection: -lig is a nominalizer applied to an ablative form

What’s wrong? • Inflectional affixes follow clitics with a lexical-like meaning • Derivational affixes follow inflectional ones, thus creating new lexemes from fullfledged inflectional forms (and not from “stems”, as expected) • Bound units and affixes may behave as (free) words and relate to lexical rather than to grammatical domain

On the other hand • Free words may behave as – or look like – grammatical markers • Unquestioned cases: auxiliaries integrated in morphological paradigms • Latin: periphrastic perfective past of middle verbs – vivit ‘lives’ ~ vixit ‘lived’ vs. moritur ‘dies’ ~ mortus est ‘died’ • Periphrastic forms fill up an obvious gap in an (otherwise coherent) inflectional paradigm and are semantically irreproachable

Less unquestioned cases • Lexical units which express “precarious” meanings and are not readily integrated into existing morphological paradigms • Modal verbs vs. inflectional “moods” – caveat emptor vs. let the buyer beware • Articles vs. case/number affixes – NB: different ways of representation in formal theories accounting for NP structure!

“Light verbs” • Persian harid purchase nešān sign harf sound sar head kardan make dādan give zadan beat ‘buy’ ‘show’ ‘talk’ ‘look into’

Light verbs • An intuitively tangible difference between “basic” (like ‘make’, ‘give’, ‘beat’, ‘bring’, etc. ) and construction-dependent meaning of a light verb which can be represented as a cluster of very abstract grammatical features (if at all) • On the other hand, derived uses of light verb are not reminiscent of typical inflectional meanings • Neither lexical nor grammatical?

“Lexical plurality” • A family of Russian constructions with the general meaning ’a lot of Ns’, ‘much N’ [Rakhilina & Lee Su Hyoun 2010] – kuča problem ‘a lot [lit. ‘heap’] of problems’ – more sčast’ja ‘a lot [lit. ‘sea’] of happiness’ – gruda del ‘a lot [lit. ‘pile’] of affairs’ –… • Non-arbitrary combinability, classifying tendencies

Lexical vs. grammatical • Semantically, “light verbs” are equally distal from core lexical and core grammatical domains • “Lexical plurality” markers are more reminiscent of core grammatical meanings, but they are much more exuberant in specific detail and appeal to (relatively) small lexical classes of nouns

Lexical vs. grammatical • The intermediate area between core lexical and core grammatical meanings is vast and heterogeneous • A useful label could be “emergent grammaticalization markers” • But can a notional border be determined?

Lexical vs. grammatical • Surprisingly little reflection in modern linguistics • Many theories do not consider the problem of lexical / grammatical distinction as central (or do not consider it at all) • Often, it is taken for granted

Hopper & Traugott 2003: 4 <…> it is usually accepted that some kind of distinction can be made in all languages between "content" words (also called "lexical items, " or "contentives"), and "function" words (also called "grammatical" words). The words example, accept, and green (i. e. , nouns, verbs, and adjectives) are examples of lexical items. Such words are used to report or describe things, actions, and qualities.

Hopper & Traugott 2003: 4 The words of, and, or, it, this, that is, prepositions, connectives, pronouns, and demonstratives, are function words. They serve to indicate relationships of nominals to each other (prepositions), to link parts of a discourse (connectives), to indicate whether entities and participants in a discourse are already identified or not (pronouns and articles), and to show whether they are close to the speaker or hearer (demonstratives).

Hopper & Traugott 2003 • Are there other “function words” beyond “prepositions, connectives, pronouns, and demonstratives”? • What about quantifiers and classifiers? • What about tense, aspect and modality markers? • What about negation and interrogative markers? • And even if the list was exhaustive, would it be of much help?

Lexical vs. grammatical • The problem is not in that “content words” and “function words” denote different things or belong to different distributional classes • The main problem is that “content word” may equally well serve as grammatical markers • The word remains the same, but we feel that “something happens” to its meaning and use – even on the earliest stages of emergent grammaticalization

Lexical vs. grammatical • “Content words” and “function words” are not necessarily different and directly opposed units having nothing in common • Diachronically, grammatical units always develop from lexical units • And more often than not, we deal with the same units synchronically • Thus, the problem of notional demarcation is crucial

Boye & Harder 2012 Kasper Boye & Peter Harder A usage-based theory of grammatical status and grammaticalization Language, 2012, 88. 1, 1 -44

Boye & Harder 2012 “So paradoxically, those who wish to claim that grammaticalization is a distinct type of change that moves across the continuum from lexicon to grammar, find themselves entangled in the selfsame problem that they brought into focus in the first place: the problem of defining in a satisfactory manner the distinction between lexical and grammatical expression. The problem simply will not go away” (2)

Boye & Harder 2012 “The central idea <…> is that grammar is constituted by expressions that by linguistic conventions are ancillary and as such discursively secondary in relation to other expressions – and that grammaticalization consists in the diachronic change that leads to such expressions. Conversely, lexical expressions are by linguistic convention potentially primary in terms of discourse prominence”

Boye & Harder 2012 “The concept of discourse prominence is understood in terms of a core idea that we regard as essentially uncontroversial: in entertaining complex mental content, there is always a priority dimension involved, so that some parts of the content are more highly prioritized than others. ”

Lexical vs. grammatical • Grammaticalization is primarily a semantic change • Other types of changes (usually attributed to grammatical units) may or may not result • Every semantic theory must account for this special type of semantic change (alongside with studying synonymy, polysemy etc. )

Acquiring grammatical status • Nonfocalizability: there is no linguistic means (as cleft constructions, focus particles, etc. ) to focalize grammatical expressions • Nonaddressability: there is no linguistic means (as confirming, denying, emphasizing, etc. ) to “take up” grammatical expressions in a subsequent discourse

Acquiring grammatical status • Grammatical markers are modifiers: they ascribe a property to a lexical item (as TAME, quantification, semantic/syntactic role, etc. ) • This property, subsequently, is ascribed discursively secondary status (in B&H terms) • A lexical item must, first, pass into “property words” class and, then, come to express a special type of property

Acquiring grammatical status • Russ. kuča ‘heap’ >> ‘many’ – ceasing to denote a physical object – beginning to denote a (quantitative) property – consequently, ceasing to be a free word – beginning to be a bound word (weakly autonomous) – preparing for a would-be loss of discourse prominence (not yet accomplished)

To conclude • Grammatical status is acquired through two important semantic processes: – A semantic shift from independent meaning to modifier meaning (expressing property) – Property meaning becomes discursively secondary – All the other features of grammatical markers result from these basic changes

To conclude • Ultimately, grammar is rooted in the lexicon • A unified theory of lexical semantics must account for basic grammatical changes (which is not always the case) • With this background, the area of emergent grammaticalization seems particularly important • It is the locus where lexical semantics interacts with grammatical status in an especially close way