Kolokace polysémních slov v psaném a mluveném projevu
In everyday communication, whether written or spoken, we rarely have problems how to resolve ambiguity resulting from the asymmetry between form and function. Those cases when one form has more than one function (or meaning), defined as homonymy or polysemy, seem complicated only in ”langue” and not in ”parole”. In ”parole”, i. e. in actual language we are provided with context, both situational and linguistic, and the context disambiguates.
Lexicographers (and not only they – also, for instance, the linguists working in the field of machine translation) are in a more complicated situation. Dictionaries are alphabetic lists of separate words which are described in linear entries. Lexicographers are mostly aware of the fact that enumeration of the distinct senses of a polysemous word is an abstraction per se. Rather than several discrete meanings there is a semantic spectrum of a polysemous word. But the role of syntactic patterns of a word in a particular sense and the role of lexical environment, i. e. of words which co–occur with the word described, usually reflect in dictionaries to a very limited extent.
According to J. M. Sinclair the reason for that is ”decoupling of lexis and syntax”.
For my research in the field of Czech monolingual lexicography I have chosen a limited set – adjectives.
Adjectives belong to the open–class items, possessing full lexical meaning, and as such are traditionally considered to be relatively autonomous, context independent. Within this word class there is no homonymy in Czech (except for a single homograph pair), therefore my paper deals only with polysemy.
In Czech, adjectives have mainly two syntactic functions: attributive and predicative. It is generally accepted that most adjectives can have both functions and only a few of them are limited either to attributive or to predicative function (pouhý, převážný vs. povděčný, nápomocný).
But what is usually not mentioned is the fact that this is not primarily the question of lexemes and that also for particular senses of a polysemous adjective either attributive or predicative syntactic pattern (or both) are characteristic. The only difference may be that the restrictions are not that absolute as with lexemes.
Semantically, adjectives in Czech are differentiated into two groups: qualitative adjectives and adjectives denoting relation, the former being further classified into two subgroups: descriptive adjectives and evaluative adjectives.
Adjectives denoting relation are often denominal, stative, characterizing nouns, expressing their permanent features. Their typical function is attributive. They can also appear in the predicate but the primacy of the attributive use is evident, their predicative function being a kind of syntactic transformation reflecting the functional sentence perspective. This is the same situation as with possessive adjectives.
When an adjective belonging to this semantic type is polysemous, it often shows the features of relational adjective in its core sense only. Any meaning extensions are inevitably followed by the transposition from the group of adjectives denoting relation to qualitative adjectives. The consequences of this are numerous. In its second sense such adjective has both syntactic functions, is mostly gradable, can form derivatives (i. e. adverbs and nouns ending in –ost), has antonyms including those derived with prefix ne–, is often expressive and functioning as a member of an idiomatic expression.
The only aspect which remains mostly unchanged are sets of nouns collocating with this adjective. This is particularly true (for instance) about adjectives denoting imaginary creatures (obří – giant, andělský – angel, trpasličí – dwarf), animals, flowers.
Let’s take for example the adjective lidský: the meaning is either concerning people or humane. In both meanings it collocates with the noun zákon ’law’. Thus we have lidské zákony ’human laws’ and ’humane laws’ but only in the second instance the adjective may become a base for derivations like: lidštější, nelidské, lidskost zákonů.
My conclusion is that all formal changes enumerated above effectively support the intuitive disambiguation of meaning and compensate partly the lack of collocational distinctions.
Qualitative adjectives, especially the subgroup expressing evaluation, can function both predicatively and attributively, are gradable, can form derivatives, have antonyms, etc.
This is a general characteristics usually given in grammars. But these features never apply to the lexeme as a whole, their occurrence differs with different senses. What is probably the most striking is the aspect of gradability. Actually, common evaluative adjectives (such as malý ’small’, velký ’big’, starý ’old’, mladý ’young’) are non–gradable in several meanings listed in dictionaries. The gradable sense which is usually not mentioned in dictionaries is that of ”having relatively more of the quality denoted by the adjective than something or someone else”. Being younger, smaller, bigger or older has nothing to do with being young, etc. The comparison introduces the notion of relativity. And of course in this meaning the adjective has only two forms: comparative and superlative while several other meanings have only the positive degree.
Also the syntactic role of qualitative adjectives differs with different senses and contributes thus to the disambiguation of meaning. Let’s have a look at the adjective přesný. We have přesný popis, výpočet, údaje ’accurate description’, ’computation’, ’data’ or přesný přístroj, hodinky ’accurate apparatus’, ’watch’. In these two senses the adjective can function both in the attribute and the predicate, collocating nouns are always inanimate but belonging to different restricted sets. This adjective also collocates with nouns denoting persons: in an attributive function it means ’meticulous’: přesný dělník ’meticulous worker’, while in a predicative function it means ’punctual’: Jan je vždy přesný. (There is one more meaning being very much of an idiomatic nature, expressing positive evaluation which is used mostly by teenagers: mít přesnou trhu, to je přesný.)
Another formal phenomenon supporting sense distinctions is valency. For instance, bohatý ’rich’ functions in syntactic patterns in both senses, but only in its second sense its valency is related to things or concepts and of course only in the predicate.
What is of crucial importance as far as the semantic features of the co–occurring noun are concerned is its animateness or inanimateness – not grammatical but rather ”natural” referring to objects in reality representing living beings. Thus, for instance, government, nation are treated as animate.
I think dictionaries should always give us explicit information whether the adjective described co–occurs with animate or inanimate nouns even in cases when changes in meaning are minor and do not have the status of separate senses.
Existing Czech dictionaries rely on this criterion only when the distinction between senses is quite evident. But there are minor changes, i. e. smutný ’sad’: We may draw a rough distinction: 1) ’showing sorrow’, 2) ’causing sorrow’. In the first sense the adjective collocates with nouns denoting living beings and with a limited set of nouns denoting devices to express sad feelings – as eyes, voice, face. There is some inconsistency in dictionaries as far as similar cases of this type are concerned (cf. smutný vs. vděčný). Sometimes they fall into one sense, sometimes they are divided into separate senses.
Another important phenomenon signalling the meaning distinctions with adjectives is a lexical set the cooccurring noun belongs to. This is of course a typical subject of corpus based studies.
I would like to mention here one criterion traditionally considered to be of great help in the lexicographic analysis of polysemous words – the criterion of synonyms. Let me point out that with synonymous adjectives we often find parallel meaning distinctions: e. g. adjectives with the core meaning ’nice to look at’ have often also the expressive meaning ’large in amount or profitable’. The criterion of synonymy is thus of limited use in the polysemy analysis.
To sum up very briefly this part of my paper: separate meanings of polysemous adjectives differ in a range of formal and semantic properties and a good dictionary should find a reasonable way how to provide us with sufficient information about that. This concerns not only a dictionary aimed at non–native speakers but also a monolingual explanatory dictionary of Czech which is used predominantly by native speakers.
In the second part of my paper I will try to answer the following question. ”Can we say that there are certain formal and semantic features of adjectives typical of written mode only and other occurring entirely in spoken mode?”
It is extremely important to answer this question in connection with lexicography. Particularly in our situation when we are approaching the design of a general corpus. In other words we could ask: Would a corpus of written texts only provide us with relevant and sufficient data for lexicographic description, or do we need a spoken corpus as well?
The answer to this question given by the corpus based lexicographic projects all over the world is yes, of course.
But let’s have a look at adjectives from this point of view. Generally, we can assume that what applies to the whole vocabulary is valid also for adjectives. And so we can presume that no adjectives are entirely excluded from either of the two modes. But many features (e. g. the absolute and relative frequency figures concerning the occurrence of lexemes, separate senses, syntactic functions, expressive meanings, idiomatic uses) – all these figures will differ being motivated not that much by either written or spoken mode of parole, but predominantly by sociolinguistic factors.
Let me make one concluding remark concerning syntactic functions of adjectives in spoken vs. written. It is obvious that one of the adjectives secondary functions, that of the head of a noun phrase, will be rare in written language and very frequent in spoken discourse. There is a whole continuum of cases ranging from a simple ellipsis of a noun due to a situational context, the other extreme being an adjective fully lexicalized as a noun. Czech dictionaries up to now have covered only the latter cases since the necessary prerequisite of a more thorough lexicographic description of this field is an access to large and relevant computer data. And the same holds for the vocabulary as a whole.
Filipec, J: Lexikálně sémantická výstavba hesla – ústřední otázka lexikografické práce. SaS, 18, 1957, s. 129–150.
Filipec, J.: K úkolům české lexikologie. SaS, 29, 1968, s. 259–276.
Filipec, J.: K otázce sémantického popisu lexikálních jednotek. SaS, 34, 1973, s. 78–84.
Filipec J. – Čermák, F.: Česká lexikologie. Praha 1985.
Sinclair, J. M. (ed.): Looking up. London 1987.
Sinclair, J. M.: Corpus, Concordance, Collocation. Oxford 1991.
R É S U M É
Článek se zabývá problematikou polysémních adjektiv a dílčími otázkami jejich lexikografického zpracování. Zatímco v běžném dorozumívání se víceznačnost slov odstraňuje intuitivně v jazykovém a situačním kontextu, lexikografové stojí pokaždé před náročným úkolem – vyabstrahovat z kontinuálního sémantického spektra jednotlivé významy a definovat je ve slovníkovém hesle. Při analýze adjektiv jsou jim oporou formální a sémantická kritéria, jako (primární) syntaktická funkce adjektiva ve větě, valence, tvoření odvozenin, stupňovatelnost, kolokační paradigma. Různé významy polysémního adjektiva se v celé řadě těchto aspektů odlišují a je úkolem dobrého výkladového slovníku, aby uživateli tuto informaci zprostředkoval.
 Cf. Filipec – Čermák (1985, p. 139).
 Sinclair (1991), p. 103: ”It is rare for a grammar to note that a certain structure is only appropriate for a particular sense of a word. … Equally, it is rare for a dictionary to note the common syntactic patterns of a word in a particular sense. … The implicit stance of a conventional dictionary is that most of the words in daily use have several meanings, and any occurrence of the word could signal anyone of the meanings. If this were actually the case, communication would be virtually impossible.”
 To restrict the material even more I left out possessive adjectives as well as shorter adjectival forms, which are formal and often getting out of use.
 Existing Czech monolingual explanatory dictionaries, e. g. the Dictionary of the Literary Czech Language (4 vol., 1960 – 1971) and the Dictionary of Literary Czech (1978) do not provide us with this information.
 It migh be interesting to note that there are some hints of transposition going in the opposite direction (i. e. from descriptive adjectives to those denoting relation). Of course this phenomenon is rather marginal. I have noticed it analysing adjectives denoting colours. Colours are frequently used as symbols of various social attitudes. In their core sense adjectives denoting colours are descriptive having most of the format features I have mentioned above. But in their symbolic use they lose most of them.
Ústav pro jazyk český AV ČR
Slovo a slovesnost, ročník 54 (1993), číslo 3, s. 216-219
Předchozí Zdena Palková: Auditiv bedingte Einschränkungen im Aufbau der mündlichen Äußerung
Následující Dana Slančová, Miloslava Sokolová: The norms of spoken communication in East Slovakia