Dėl gramatinių funkcijų skirstymo

PASTABOS | 107 ON THE CLASSIFICATION OF GRAMMATICAL FUNCTIONS The review of Studies in Syntactic Relations, written by Prof. Vytautas Ambrazas and published in 2005, had already delighted the authors of the Studies in Grammar series, launched in 2003. This issue of ALL publishes a review by the same author of the third volume of Studies in Grammar, Studies in Grammatical Functions. It has prompted several thoughts that I would like to share. It is gratifying that this collection of articles was not simply ignored as an event irrelevant to Lithuanian grammar, but judging by the review, it does not seem to have sparked a dialogue. Although the reviewer acknowledges that the ideas raised in the collection should be considered thoroughly by researchers of the Lithuanian language, he suggests doing so only on the basis of the concepts used in previous grammars of Lithuanian, especially the definitions of types of syntactic relations presented in those grammars. For, “there is probably no need to prove that the aforementioned types of relations are syntactic, rather than morphological, semantic, or otherwise.” If this does not need to be proved, then apparently it is an axiom. Yet an axiom may be accepted or rejected. Alongside Euclidean geometry there are other geometries not based on its axioms. Of course, no one attacks this geometry for relying on certain axioms, but anyone who found internal contradictions in it would have the right to point them out. This is precisely what the authors of Studies in Grammar attempted to do with respect to The Grammar of Contemporary Lithuanian, drawing attention to internal contradictions in some of the matters presented there. However, according to Prof. Ambrazas, “This approach to syntactic relations is rejected by the authors a priori, for some reason on the assumption that the concept of a bidirectional syntactic relation and mutual linkage [...] is unnecessary and arose from confusing syntactic and morphosyntactic phenomena.” But it is not rejected a priori, and certainly not “for some reason.” In the footnotes, the authors argue in detail why it is rejected, pointing out that the concept of bidirectional dependency leads to irresolvable contradictions. The reviewer says nothing about any of these arguments, as though they had never been made. Of course, there was apparently no need to discuss them, since the text was conceived as a review of Studies in Grammar. Nevertheless, using the word “for some reason” here is inappropriate. The reviewer goes on to write: “In fact, there is no confusion here; it is an obvious fact that syntactic relations, and thus mutual linkage as well, have certain morphosyntactic markers.” No one denies that syntactic relations can have morphosyntactic markers. Precisely, they “can have” them, since traditional grammar also recognizes “adjunction” in cases where the dependent word has no marker of dependency, e.g. Jonas atbėgo greitai. And it is precisely here that a certain difficulty lies, one that the reviewer consistently leaves unmentioned. Traditional grammar uses two entirely different heuristic methods for identifying syntactic dependency. On the one hand, agreement in sentences such as Mačiau Joną linksmą provides grounds for claiming that the word linksmą is syntactically dependent on the word Joną. On the other hand, in sentences such as Jonas atbėgo greitai, where morphosyntactic criteria cannot be relied upon, it is assumed that greitai depends on atbėgo. Thus, a morphosyntactic criterion is not always used when determining whether a syntactic relation exists. Since this dependency is neither government nor agreement, it is described as adjunction. This is precisely where the danger lies: the heuristic methods are contradictory. If it is permissible to identify syntactic dependency even in cases where there is no morphosyntactic marker, then, for example, it is permissible to claim that in the sentence Laikyk butelį stačiai the word stačiai is adjoined to the word butelį. For how does this sentence differ from Laikyk butelį stačią, where a dependency of this type on the object is recognized? Only in part of speech—the first sentence uses an adverb. Anyone making such a claim would simply be relying on 108 | PASTABOS the criteria proposed by Prof. Ambrazas himself. He would be doing nothing different from those who assume an adjunction relation between an adverb and a verb in sentences such as Jonas atbėgo greitai, Laikyk butelį stačiai. It would then follow that the syntactic relation is double both in the sentence Laikyk butelį stačiai, and in the sentence Laikyk butelį stačia. What does it matter that the authors of The Grammar of Contemporary Lithuanian see no need, in the first case, to identify adjunction between butelį ir stačiai? What matters is only that this can be done on the basis of the criteria proposed by this grammar. It follows, then, that the claim that the syntactic relation is double in the sentence Laikyk butelį stačiai is just as correct as the claim that it is single. A student who identifies a double syntactic relation in this sentence can be marked down, but it cannot be proved, on the basis of the theory in The Grammar of Contemporary Lithuanian, that the student is wrong. No sophistry can eliminate the flaw in this system, which is based on contradictory heuristic methods. How obvious, in fact, are the facts cited by the reviewer? Let us quote once again: “it is an obvious fact that syntactic relations, and thus mutual linkage as well, have certain morphosyntactic markers.” But following the landmark article published by Joanna Nichols in 1986, it is clear that the directions of syntactic and morphological dependency can be opposite; this phenomenon is called “head marking” (head marking). Thus, a verb governing a noun (the head) may agree with it; in many of the world’s languages, verbs agree not only with subjects but also with direct and indirect objects. It is hardly likely that the concept of “head marking” is unfamiliar to Prof. Ambrazas, but apparently he considers the spread of this idea in Lithuanian linguistics undesirable. If it is accepted, mutual linkage becomes unnecessary. Anyone may reject the concept, but in 2005 its denial is no longer an “obvious” fact. Without a word about the arguments on which the authors of Studies in Grammar base their proposal to change the conception of syntactic relations, the reviewer devotes his efforts to proving that these authors contradict themselves, acknowledging syntactic dependency in one place and rejecting it elsewhere. It is apparently not difficult to do this by taking individual sentences out of context. For the authors of Studies in Grammar discuss various types of dependency—semantic, syntactic, and morphological. In the authors’ view, distinguishing them is important so that the conceptual apparatus is precise and the theory consistent. Restoring the truth by examining each quotation and showing how and for what purpose the quoted sentence was taken out of context and distorted would be a long and tedious process. I will give just one example. The authors of Studies in Grammar are reproached for trying to “deny the very syntactic relation between the object and the predicative complement in sentences such as Jis parvertė brolį aukštielninką [...] arba Mačiau jį linksmą,” and it is added: “Yet elsewhere they propose ‘to associate the term secondary predicative specifically with [...] secondary predicates that have the relevant morphosyntactic features.’” Correct: this is what Studies in Grammar says, and the two claims do not contradict each other. For the authors of Studies in Grammar specifically intended to show that the distinction between predicative complements and adjuncts is not syntactic but morphological. This is precisely what the reviewer leaves unmentioned. Anyone wishing to learn why it is better to abandon the double syntactic relation of the predicative complement can turn to page 490 of The Grammar of Contemporary Lithuanian, where, alongside agreeing predicative complements, non-agreeing ones are also distinguished. But don’t predicative complements differ from adjuncts precisely through agreement (an adjunct is adjoined to the verb, while a predicative complement is also agreed with the subject or object)? It is precisely agreement in the sentence Mačiau jį linksmą, according to Prof. Ambrazas, that indicates a syntactic dependency between the predicative complement and the object, which allows syntactic PASTABOS 109 to describe the relation as twofold. If agreement is included in the definition of a predicative modifier, then what on earth is a “non-agreeing predicative modifier”? Since it is better to avoid such a contradictio in adiecto, Gramatikos darbuose it is proposed to acknowledge that the distinction between predicative modifiers and adverbials is morphosyntactic, not syntactic. What is actually wrong here? The reviewer accuses the authors of Gramatikos darbų of conflating criteria, when in fact they merely point out that it is useful to distinguish syntactic dependency from morphological dependency. When the two are forcibly lumped together, the result is, for example, “non-agreeing predicative modifiers.” Where do the criteria actually “cross,” and where do they not? One more remark on the very notion of “contradiction.” In the reviewer’s opinion, many of the points presented in Gramatikos darbuose are contradictory. For instance, the statements about predicative modifiers just quoted are supposedly contradictory. The notion of an obligatory adverbial is “obviously contradictory.” Why this notion is “obviously” contradictory is unclear, since the author of these lines has already explained in the work under review what the chosen term is based on and why it is not contradictory (the reviewer does not inform the reader how the authors define the notion of an obligatory adverbial; thus, its contradictoriness is presented as an article of faith, much like syntactic relations that supposedly need no proof). Perhaps, if the reviewer’s criteria for “contradiction” were understood, it would be easier for the reader to see why the notion of a “non-agreeing predicative modifier,” introduced immediately after defining a predicative modifier in terms of agreement with the subject or object, is not “obviously contradictory.” I would argue, on the contrary, that the notion of a “non-agreeing predicative modifier” reveals an internal contradiction, since it contradicts the definition given several pages earlier in the same chapter, whereas the contradictions found in Gramatikos darbuose are contradictory only from the reviewer’s perspective. The purely morphological distinction between predicative modifiers and adverbials is contradictory only in that it conflicts with the definitions of syntactic relations presented in Dabartinės lietuvių kalbos gramatikoje, whose correctness “needs no proof.” Finally, the last sentence of the review is also striking: “Thus, we must agree with the authors themselves that ‘they have not managed to satisfactorily resolve the problems associated with defining and describing the functions mentioned.’ Let us hope that this can be done in the future.” In fact, the authors of the volume under review never expected to solve all the problems associated with describing grammatical functions. No school of linguistics has yet fully succeeded in doing so, and there is ample reason to doubt that all the questions that have troubled linguists for decades will ever be definitively answered. Every linguist probably shares such doubts, which makes both Prof. Ambrazas’s remark and the assumption behind it all the more surprising: that failing to answer every question and reveal the ultimate truth of grammar (God’s truth) is solely a problem for the authors of Gramatikos darbų, one that could easily be remedied by returning to the refuge of traditional grammar. The authors of Gramatikos darbų do not expect to lure linguists away from this refuge if they feel safe there. But those who think it can sometimes be healthy to question and critically reflect on truths that supposedly need no proof may find in Gramatikos darbuose not final answers, perhaps, but at least interesting questions that Prof. Ambrazas unfortunately chose to avoid in his review. It is a great pity that dialogue is being avoided, since it could be interesting. Axel Holvoet Institute of the Lithuanian Language P. Vileišio g. 5, 10308 Vilnius, Lithuania