Skip to document
Back
Go to original Localized version was produced by AI. It may contain inaccuracies.
Preparing document pages…
1 short passage of this localized version is shown in the original language because it could not be localized reliably.

DAIVA ŠVEIKAUSKIENĖ

Institute of the Lithuanian Language

ORCID id: https://orcid.org/0000-0002-7570-9132

RESEARCH AREAS: computational linguistics, Lithuanian syntax and morphology.

DOI: https://doi.org/10.35321/bkalba.2024.97.10

GOVERNMENT OR ADJUNCTION: THE NON-AGREEING ATTRIBUTE EXPRESSED BY THE GENITIVE AND THE PREPOSITIONAL CONSTRUCTION

KEYWORDS: syntactic relations, government, adjunction, genitive, non-agreeing attribute, prepositional construction.

ABSTRACT

The article reviews the assessments of the subordinating relation presented in the scholarly literature. Other authors’ views on government and adjunction relations are examined in greater detail, especially in disputed cases. The author draws attention to the relation between a noun in the genitive case functioning as a non-agreeing attribute and the word it modifies. Cases in which this case is governed or adjoined are discussed separately, and the question of syntactic relations in prepositional constructions is also considered.

ABSTRACT

The article reviews the estimation presented in the scientific literature regarding the subordinating relation in word combinations. The opinion of other authors on the relationships of control and adjoining are examined in more detail, especially in contested cases. The author draws attention to the relationship of genitive case of the noun, which is a non-agreeing attribute, with the head word, and also discusses cases of control and adjoining the genitive of the noun. Prepositional constructions are also discussed separately in aspect of control and adjoining.

INTRODUCTION

When language is computerized, problems sometimes arise because grammars present differing assessments of the same linguistic phenomenon. When work began on the syntactic component of the Lithuanian Language Grammar Information System (LIGIS), it became clear that gram-

DAIVA ŠVEIKAUSKIENĖ. Valdymas ar šliejimas: kilmininku reiškiamas nederinamas pažyminys ir prielinksninė konstrukcija

209

mars and various works on syntax assess some cases of syntactic relations differently. Here, it would be useful to examine more closely the relation between a non-agreeing attribute and the word it modifies. It is fairly common to regard it as government, but this proved to be rather inconvenient when processing Lithuanian by computer. In computerizing Lithuanian syntax, valency studies are taken into account; these focus on “the general possibilities of word combinability, reflecting both syntactic (cf. a very good proposal, but not *very proposal) and logical-semantic or lexical word combinability, cf. The father’s hat was hanging on a nail, but not *The child’s curiosity was hanging on a nail” (Babickienė, Venckutė 2013: 169). Valency, as is well known, is a feature determined by a word’s meaning, whereby it opens a certain number of empty slots filled in a sentence (or word combination) by words and their forms, or their equivalents, with particular meanings (e.g., Sližienė 1994: 16; Vaičiulytė-Semėnienė 2023: 182). Valency dictionaries and other works on word valency provide information about the government of verbs (e.g., Sližienė 1994, 1998, 2004), adjectives (e.g., Vaičiulytė-Semėnienė 2001), and some deverbal nouns (e.g., Vaičiulytė-Semėnienė 2023), but they do not indicate that a non-agreeing attribute is governed. In computerizing language, the government of certain words (nouns, adverbs, numerals) can be identified on the basis of semantic features of the word, but a governed non-agreeing attribute cannot be detected by any means available to a computer.

Therefore, the aim of this article is to examine more closely the relation between a non-agreeing attribute and the word it modifies, that is, to consider whether it is government or adjacency, as well as to draw attention to the syntactic relations of prepositional constructions and indicate which solution will be adopted in the Lithuanian Language Grammar Information System. The article is a critical analysis employing the deductive method.

1. TYPES OF SYNTACTIC RELATIONS IN VARIOUS PUBLICATIONS

When discussing the relations among the words that make up a sentence, semantic and syntactic relations are usually distinguished; the latter are generally divided into two main types, e.g.:

The words that make up a sentence have certain connections with one another. Relations between words are usually divided into two categories: 1) internal, or semantic, relations, and 2) external (formal), or grammatical (syntactic), relations. Internal relations between words are objective, naturally existing relationships. <...> In traditional linguistics

210

BENDRINĖ KALBA | 97

The external (formal) relation is called a grammatical relation. In linguistics, the term syntactic relations is often used for it. Syntactic relations may be of two types: 1) coordinate; 2) subordinate1 (Aliyeva 2014: 41–42).

Perhaps the greatest number of conflicting opinions concern subordinating relations. In Lithuanian syntax, three types of subordination are generally distinguished according to the formal expression of the relation: agreement, government, and adjunction (see Babickienė, Venckutė 2013: 171). There seem to be no major disagreements about agreement. This relation is described in very similar terms in various works: “Agreement is the adaptation of the form of a dependent word to the head word according to a morphological category. The dependent word’s agreeing form acquires the case, number, and gender markers characteristic of the head form” (Ambrazas 1986: 8); “Agreement is the adjustment of the form of a dependent constituent to the form of the head constituent” (LKG 1976: 11); “Agreement is a type of subordination in which the gender, number, and case of the head word determine the corresponding form of the dependent word” (ŽS).

Gal tik naujausiuose, į Vakarų lingvistiką orientuotuose darbuose paabejojama dėl linksnio derinimo; aiškiu laikomas tik skaičiaus ir giminės derinimas (Holvoet 2009: 110). Štai aptardamas pavyzdį noriu skanaus pyrago Axelis Holvoetas mini, kad pagal sandų teoriją būtų teigiama, jog linksnį gauna visas sandas skanaus pyrago: „Priklausomybių gramatikoje logiška manyti, kad būdvardis linksnio požymį gauna iš daiktavardžio, o šis – iš veiksmažodžio. Sandų struktūra paremtoje sintaksėje būtų teigiama, kad skanaus pyrago sudaro sandą, kuris kaip visuma gauna linksnį iš veiksmažodžio“ (Holvoet 2009: 111). Tačiau šio darbo autorės jau ir anksčiau atkreiptas dėmesys, kad sintaksės aprašymo būdas ar sintaksinės analizės pobūdis, puikiai tinkantis vienos kalbos sintaksinei sandarai perteikti, nebūtinai perteikia kitos kalbos sintaksines ypatybes (Šveikauskienė 2013). Vitas Labutis taip pat yra atkreipęs dėmesį į tuos skirtumus, t. y. rašęs, kad amerikiečių sukurtos frazių gramatikos pagrindinis principas aprašyti kalbą atsižvelgiant tik į linijinę sakinio struktūrą, t. y. pagal žodžio apsuptį jungiant žodžių grupes į sandus, lietuvių kalbai netinka: „Neturinčioje griež-

1 “Words, which are combined and constitute a sentence, should have some link relations between each other. Usually, the relations between the words are divided into two parts: 1) the internal or the meaning relations and 2) external (formal) or grammatical (syntactical) relations. The internal relations between the words are the objective relations which exist in the nature. <...> In traditional linguistics the external (formal) relationship is called the grammatical connection. It is often called syntactic relations in linguistics. Syntactic relationship is expressed in two forms: 1) Coordinating relations; 2) Subordinating relations” (Aliyeva 2014: 41–42).

DAIVA ŠVEIKAUSKIENĖ. Valdymas ar šliejimas: kilmininku reiškiamas nederinamas pažyminys ir prielinksninė konstrukcija

211

Dependency grammar, based on valency theory (Tesnière 1959), was developed for the inflectional languages of Europe, since their structure differs in principle from that of English. From the perspective of dependency grammar, “<...> the task of syntax is not so much to group words into phrases as to establish direct relations between the words themselves” (Kay, Gawron, Norvig 1994: 53). Thus, dependency grammar became widespread because it was much better suited to the inflectional European languages, which have free word order in the sentence, than phrase grammar, which operates with constituents:

It is no coincidence that dependency grammar was most suitable for researchers of languages in which word order is relatively free and the notion of a phrase seems less important, because the main element of a word combination and its dependent elements are not always arranged in the same order in a sentence, and may not even be adjacent to one another2 (Kay, Gawron, Norvig 1994: 55).

Linguists in other countries also write that the nature of syntactic analysis can vary greatly and depends on the language. For example, it is noted that in Russian much is determined by case and verb forms, word order may also provide some information, whereas in English almost everything is determined by word order, and variants of noun and verb forms can contribute little to syntactic analysis (Henisz-Dostert, Macdonald, Zarechnak 1979: 116). In English, syntactic relations are determined by endings only in very rare cases, e.g., I know Danny and Toni knows me and I know Danny and Toni know me (Winograd 1983: 136); therefore, theories oriented primarily toward English make no provision for extracting information from word endings, which have a decisive influence on the syntactic analysis of Lithuanian sentences.

1.1. Description of government

2 “It is no accident that dependency grammar has appealed most to students of languages with relatively free word order where the notion of a phrase appears less important because a head and its dependents may not always appear in the same order or even adjacent to one another” (Kay, Gawron, Norvig 1994: 55).

212

BENDRINĖ KALBA | 97

1.1. Description of government

In some earlier works by Lithuanian authors, only the government characteristic of verbs is highlighted, e.g.: “It is customary to classify verbal combinations according to the nature of the governed cases or cases with prepositions” (Grenda 1964: 137). Later, government characteristic of other parts of speech is also discussed more broadly:

Words from various parts of speech can govern: verbs (build a fire, fear a wolf, say to a friend, take an interest in music, sit in a tent, be angry with one’s brother, walk through a forest, get along with a neighbor), nouns (a dog’s fight with a wolf, a pot without handles, a sack of potatoes, a tree branch...), some adjectives and adverbs (full of water, friendly to children, remarkable in that, much trouble...), numerals (a hundred books, a pair of horses, ten of you...) <...> (Labutis 1998: 34).

At first glance, there seems to be general agreement on the definition of government: “Government is the selection of an inflected form (usually a different one) of the dependent word, determined by the valency of the head word” (Ambrazas 1986: 8); “<...> when the form of the dependent word is determined by the property of the main constituent of taking a particular case, <...> the two constituents are linked by government (Babickienė, Venckutė 2013: 172); “The government of words in a sentence is a relation in which one word takes the form (case) required by another word on the basis of meaning” (Žiugžda 1960: 16); “In government, one word imposes on another a certain form that is not characteristic of the latter itself” (Holvoet, Judžentis 2003: 18).

However, opinions differ regarding the treatment of individual cases from the perspective of government or adjoining. For instance, Vitas Labutis (1998: 35) gives the locative as an example when describing verb government: to sit in a tent. The three-volume Grammar of the Lithuanian Language (hereafter LKG) states that “one can speak of government of the locative only with certain reservations” (LKG 1976: 12), and verb combinations with the locative are discussed under adjoining. Aldona Paulauskienė, however, argues that “the locative is certainly not a governed case, because its meaning (like that of an adverb) does not depend at all on the verb with which it forms a combination” (Paulauskienė 2002: 25). As an argument against the locative being governed, the author notes that when this case in a verbal phrase is replaced by an adverb, e.g., there, the result is adjoining.

At school, students are taught to identify syntactic relations by asking a question. That to live where? is not government would be shown by a sentence such as: Then they were lovely and lai-

DAIVA ŠVEIKAUSKIENĖ. Valdymas ar šliejimas: kilmininkų reiškiamas nederinamasis pažyminys ir prielinksninė konstrukcija

213

lived happily. However, to want cannot be used this way. *They really wanted – without a complement, the meaning of the sentence is not sufficiently clear; information is missing. The reason for the complement meaning is “the semantic insufficiency of the word” (see Vaskelaitė 2009: 81). One cannot say that to live is semantically insufficient. The sentence given, Afterwards they lived beautifully and happily, is completely clear. Thus, such examples would lead one to conclude that the verb to want governs the genitive, whereas the verb to live does not govern the locative.

1.2. Description of Adjoining

Narmin Aliyeva gives the most rudimentary definition of adjoining: “<...> if the relationship between two words is neither agreement nor government, then it is adjoining”3 (Aliyeva 2014: 43). In LKG, adjoining is “the kind of relationship in which a dependent constituent attaches to a head constituent without being required by the latter; it is formally unmarked and is connected to the head constituent only logically” (LKG 1976: 12). Other authors describe this relationship between words similarly: “<...> adjoining is a way of expressing a subordinating relation in which a word or one of its forms does not formally indicate a syntagmatic relation; this is determined solely by meaning” (Labutis 1998: 36); “It is a way of attachment unrelated to changes in form (formally unmarked), based solely on semantic relations between words and not determined by the valency of the head constituent. In this way, invariable words or invariable forms of individual words are generally adjoined: adverbs (write correctly, very far, black and white, late at night), infinitives (began to write, the right to learn, a desire to please), and converb forms (got up at dawn, came home after everyone had gone to bed)” (Bučienė 2014: 30). It is further noted that adjoining “is usually a weak subordinating relation” (Bučienė 2014: 31), since “an adjoined word is not necessary to the meaning of the head constituent (fell asleep while reading, reads aloud, I thought it over after waking up)” (Bučienė 2014: 31). Government, incidentally, is also divided into weak and strong, e.g.:

A distinction is made between obligatory, or strong, government, in which verbs necessarily require a particular case form that complements their meaning, and non-obligatory, or weak, government, in which a particular case form is not necessary for verbs but is merely optional. <...> However, it is not possible to draw a strict boundary between strong and weak government, or between weak government and adjoining (LKG 1976: 12).

3 “<...> if there is no concordance and control relation between two words, then it is the adjoining relation“ (Aliyeva 2014: 43).

214

BENDRINĖ KALBA | 97

A. Paulauskienė does, however, draw attention to the overly broad conception of government and argues that only instances of strong government should be considered government, while weak government, in her view, should be considered adjoining:

Clearly, the cases answering the questions “of what?”, “to whom?”, “what?”, and “with what?” are governed: they are directly required by the verb functioning as the predicate, and they are complements. But what should we call the circumstantial, rather than objective, relation between a verb and a case? After all, adverbs expressing circumstances are adjoined <...>. When a noun itself denotes not an object but a circumstance, some cases become adverbial. And it is not the governed cases that become adverbial, but the adjoined cases, especially the locative. Perhaps this is why various degrees of strength of government (strong and weak) are distinguished. The locative is certainly not a governed case, since its meaning (like that of an adverb) does not depend at all on the verb with which it forms a phrase (Paulauskienė 2002: 25).

Essentially the same view is expressed in the most recent literature: “From the perspective of semantic relations, weak government is no longer government, since the dependent word is not required” (Judžentis 2022: 148).

1.3. Comparison of Models for Describing Subordinating Relations

As theories of syntax developed abroad reveal, they operate not with the concepts of government and adjoining, but with those of complementation and modification. Complementation and modification are distinguished as two types of syntactic relation (syntactic dependency). This is also evident from the description given by Christina Biava, who defended her dissertation in theoretical and applied linguistics at the University of Illinois in 1991: she calls elements that complement the meaning of words in a sentence and are obligatory complements, while elements that are not obligatory in a sentence are considered modifiers. She also provides the following information: “Modifiers are usually adverbs and adjectival words. Complements are usually nouns or pronouns”4 (Biava 2020). According to this understanding, modification should encompass agreement and adjoining as distinguished in Lithuanian syntax, while the principal criterion for distinguishing complementation from modification is whether or not an element is required as a complement. This criterion is also mentioned in works on Lithuanian linguistics: “In foreign linguistics, syntactic relations are generally defined and classified according to a single criterion—whether or not the head constituent requires the dependent” (Vaskelaitė 2003: 145).

4 “Modifiers are typically adverbials and adjectives. Complements are typically nouns / pronouns” (Biava 2020).

DAIVA ŠVEIKAUSKIENĖ. Valdymas ar šliejimas: kilmininkų reiškiamas nederinamasis pažyminys ir prielinksninė konstrukcija

215

In the most recent works on Lithuanian syntax, a model of describing dependency relations that accounts for complementation and modification is also applied to represent Lithuanian dependency relations: “<...> two types of syntactic relations (syntactic dependencies) can be distinguished: complementation (English complementation) and modification (English modification) <...>. This distinction between syntactic relations can also be formulated in terms of syntactic requirement: a complement is a syntactically required adjunct, whereas a modifier is not required” (Holvoet 2009: 72). Examples with an adjective and an adverb are provided to illustrate modification:

<...> the adjective interesting MODIFIES the noun book. <...> In the phrase interesting book, the semantic relation between the words is determined by the modifier interesting; the meaning of the noun itself, book, does not require the expression of a property, but if a property is expressed, it is clear that it is expressed by a modifier. A similar relation can be observed in the phrase tells an interesting story, where the adverb interestingly characterizes the telling, but the meaning of the verb itself, to tell, does not include or require such a characterization (Holvoet, Judžentis 2003: 14–15).

Before this model was introduced, it was customary in Lithuanian syntax to distinguish three types of dependency, namely agreement, government, and adjunction. As the description of agreement, government, and adjunction presented here shows, and as summarized in Ramunė Vaskelaitė’s dissertation, two criteria were applied to distinguish them: “LKG [Grammar of the Lithuanian Language] presents a model of dependency relations based on these two criteria: 1) the criterion of requiring a dependent constituent on the basis of the meaning of the head constituent; 2) the criterion of the formal expression of the dependent constituent” (Vaskelaitė 2009: 101).

This model for classifying dependency relations could be represented in a diagram (see Fig. 1). Agreement and government share one feature (marked in green in the diagram): the complement is expressed through forms. They differ in that with government the complement is required, whereas with agreement it is not. Agreement and adjunction share one feature (marked in purple): in both cases, the complement is not obligatory. Only with agreement is the relation expressed through forms; with adjunction, it is not. The model for classifying dependency relations applied in the most recent works on Lithuanian syntax posits complementation and modification as two types of dependency. It is shown in Fig. 2.

One might ask which model is better suited to revealing the syntactic structure of Lithuanian. In English, there are almost no inflectional endings, so Lithuanian

216

BENDRINĖ KALBA | 97

GOVERNMENT

FIG. 1. Classification of dependency relations encompassing agreement, government, and adjunction

FIG. 2. Classification of dependency relations encompassing complementation and modification

There is also no agreement between nouns and adjectives characteristic of Lithuanian (see Fig. 2). Even in the genitive or plural, when the noun takes the endings –’s, –s’ or –s, the adjective modifying it receives no ending (e.g., big dog and big dogs). Agreement (English agreement) is sometimes used here to refer to what some works on Lithuanian syntax understand as correlation (e.g., Labutis 1998: 38, DLKG 1997: 478), that is, the selection of the predicate form according to the subject expressed by a noun, as in the examples given, Danny and Toni know me and Toni knows me. “The relation between the verb and its nominal subject is often described as a system of agreement, in which certain noun forms correlate with specific verb forms”5 (Bowers 2001: 25). Such correlation is not equivalent to the term agreement as used in Lithuanian syntax. In Lithuanian academic language

5 „The relation between a verb and its subject nominal, for example, is frequently marked by an agreement system in which specific morphological forms of the nominal correlate with specific forms of the verb“ (Bowers 2001: 25).

DAIVA ŠVEIKAUSKIENĖ. Valdymas ar šliejimas: kilmininku reiškiamas nederinamasis pažyminys ir prielinksninė konstrukcija

217

The grammar states: “Words that have the same grammatical categories—gender, number, and case—are joined by agreement” (LKG 1976: 11).

When syntactic relations between words are identified in English, word order in the sentence is essential. In Lithuanian, word order carries almost no syntactic information, and direct syntactic relations in the sentence are indicated mainly by word endings, regardless of the position those words occupy in the sentence. Therefore, if modification is taken to include both adjunction and agreement, certain syntactic data conveyed by endings would be disregarded—data that show which word agrees with which, or, more precisely, that make it possible to distinguish agreeing attributes from adverbials.

Nevertheless, a classification of the subordinating relation that posits complementation and modification as two types does not exclude the concept of agreement altogether. It is merely proposed that agreement be regarded here as a morphosyntactic, rather than a syntactic, relation, and thus assigned to another level—that of morphosyntax. In other words, the proposal is to distinguish syntactic relations from morphosyntactic phenomena, without conflating them (Holvoet, Judžentis 2003: 12), and to distinguish the concepts of “syntactic and morphosyntactic dependency” from each other (Holvoet 2009: 118). The term morphosyntax, however, has appeared only in these works—that is, works proposing the application of two types of subordinating relation in Lithuanian syntax: complementation and modification. The term is not mentioned in earlier works on Lithuanian syntax. The impetus or rationale for the new description of subordinating relations in Lithuanian syntax, presented in the Lietuvių kalbos gramatikos darbuose, may perhaps be found in the preface to the first volume in this series, Sintaksinių ryšių tyrimai: “The aim of the series is to encourage discussion of issues in the theoretical grammar of Lithuanian” (Holvoet, Judžentis 2003: 7).

2. THE GENITIVE OF THE NOUN

There is disagreement about the relation between a non-agreeing attribute expressed by the genitive and the word it modifies. Morphologically, it is expressed by the genitive of a noun or pronoun, but linguists disagree about the syntactic relation between them: some maintain that it is government, others that it is adjunction.

This article distinguishes two cases: in one, the relation of the genitive as a dependent word to the head word (expressed by another noun) is government; in the other, it is adjunction. Arguments will be presented to support this distinction.

218

BENDRINĖ KALBA | 97

2.1. The case of government

When V. Labutis gives examples of noun government, he includes the word combinations maišas bulvių and medžio šaka (Labutis 1998: 34), although in the preceding paragraph he has stated that “<...> government is not as purely formal a way of expressing relations between words as agreement. To some extent, government is connected with the semantic and syntactic function expressed by case form, and also depends, to a greater or lesser degree, on the lexical meaning of the governing word” (Labutis 1998: 34). This is indeed the case in the word combination maišas bulvių: the word maišas, like other nouns denoting containers or measures, requires the genitive as a complement. But this certainly cannot be said of the word combination medžio šaka. The formulation of adjunction given, for example, by Bronius Kalinauskas is more suitable for describing this word combination: “Adjunction of words is the joining of words based solely on semantic relations and unrelated to any change in form” (Kalinauskas 1963: 9). B. Kalinauskas also mentions another feature of adjunction: “Some linguists consider what we call weak government to be the adjunction of case or prepositional constructions, because the governed words here no longer answer questions associated with object cases, but questions characteristic of adverbials or attributes (when did it happen? – in summer, what kind of ring? – gold, and so on)” (Kalinauskas 1972: 4). The same applies here: the word medžio answers an attributive question—what kind of? whose branch?—not a branch of what?, as in to want an apple – to want what?—which would be government, as would a sack of potatoes – a sack of what?, not *what kind of sack, whose sack?*

LKG also identifies government in word combinations such as tėvo knyga, draugo žodžiai, aukso žiedas (LKG 1976: 16). However, both in the discussion of syntactic relations (LKG 1976: 16) and in the description of noun phrases, it is emphasized that the genitive is required by certain nouns as a complement to their lexical meaning:

The genitive of a noun is governed by nouns that require the genitive as a complement to their lexical meaning. These include nouns denoting an indefinite quantity, e.g., truputis, dalis, mažuma, daugybė, in combinations with the genitive of concrete or abstract nouns; nouns denoting units of measure, e.g., lիտras, metras, kilogramas, tona, sauja, gabalas, in combinations with the genitive of nouns denoting materials; nouns denoting vessels or premises, e.g., stiklinė, puodelis, maišas, statinė, butelis, in combinations with the genitive of nouns denoting the items sought; nouns denoting a collection or accumulation of homogeneous things, e.g., būrys, pulkas, grupė, krūva, in combinations with the genitive of nouns denoting people, animals, or things (LKG: 1976: 202–203).

DAIVA ŠVEIKAUSKIENĖ. Valdymas ar šliejimas: kilmininku reiškiamas nederinamas pažyminys ir prielinksninė konstrukcija

219

Adjectives can also govern the genitive, e.g., worthy of what? – attention, respect, and adverbs expressing measure, e.g., many problems. Numerals also govern the genitive of nouns:

Cardinal numerals from eleven to nineteen, as well as numerals denoting tens, hundreds, thousands, millions, etc., and all collective numerals, with the possible exception of one, govern the object genitive, e.g.: ten rubles, eleven days, one hundred pupils, a thousand inhabitants, a group of four horses. In these phrases, the genitive is necessary to complete the lexical meaning of the numeral. It denotes the object whose quantity is indicated by the numeral (LKG 1976: 15–16).

When a noun’s genitive is governed by a part of speech other than a verb, that genitive usually follows the word that governs it. This position often also indicates differences in the meanings of word combinations: a box of sweets (it may be empty; an attributive relation) and a box of sweets (it necessarily contains sweets; government). This is one of those very rare cases in Lithuanian where word order matters.

One source on the valency of adjectives states: “It is inappropriate to include neuter adjective forms in the paradigm of forms of adjectival phrases, since the valency properties of this adjective form most often differ from those of the adjective” (PŽJ). This can be agreed with only in part. Adjectival government is so strong that this claim does not apply to it; compare: (the bucket is full of water and here everything is full of water, rubbish. The neuter form full governs the same genitive as its base form—the masculine singular nominative full.

2.2. The case of adjoining

The relations expressed by the genitive as a non-agreeing attribute are also assessed differently. In LKG, as already mentioned, they are treated as government: “A distinct group consists of nominal phrases with the genitive of nouns or words used nominally, e.g., father’s book, a friend’s words, a gold ring, our home. In these phrases, the genitive governed by a noun is a definite case” (LKG 1976: 16). A. Paulauskienė is surprised by this: “It is interesting that the descriptive genitive of a non-agreeing attribute, which expresses not objective but attributive relations and is moreover often used synonymously with agreeing adjectives (cf. a gold ring and a golden ring, copper gates and copper gates, evening news and evening news), is classified among governed words” (Paulauskienė 2002: 25). The government analysis of this genitive is doubted by—

220

BENDRINĖ KALBA | 97

linguists in other countries as well. For example, in her book “Noun Valency” (Noun Valency), Olga Spevak writes: “The governing word does not, by virtue of its meaning, require a genitive denoting material, origin, value, or composition6 (Spevak 2014: 186). Thus, the author argues that in phrases such as ivory, made of ivory (as well as a gold ring, a ring made of gold), the modified word does not require a genitive to complete its meaning.

The most recent literature on the issue of dependency relations is likewise oriented toward the criterion of requirement: on this basis, government (complementation) and characterization (modification) are distinguished:

The author understands syntactic relations as a phenomenon at the level of syntax based on semantic relations and most often expressed by morphological forms. Two types of dependency are distinguished—government (complementation) and characterization (modification). Government is a syntactic relation in which one word requires another because of its meaning. A relation in which no complement is necessary to a word, either semantically or syntactically, but is nevertheless syntactically connected with that word, is called characterization (Judžentis 2022: 135).

Although only two types of dependency relations are distinguished here, it is clearly stated that if the head word does not semantically require a particular complement, the dependent word cannot be governed. A. Paulauskienė also directly argues that, according to the criterion of requirement, the relation between a non-agreeing attribute and its modified word is more naturally viewed as adjoining:

The adjoining relation in the phrase the roof of a house seems much more natural; it narrows the meaning of the noun roof (as does every word with an attributive meaning), indicating to whom it belongs (a possessive relation), whereas the roof itself does not require such a description at all. E.g.: A snow-covered r o o f is slippery; climbing on it is dangerous (Paulauskienė 2002: 25).

Other attributive relations, for example the relation between an agreeing attribute and its modified word, are not government either, because a noun does not generally require an attribute—whether agreeing or non-agreeing—by virtue of its meaning. This is also stated about the agreeing attribute in publications presenting a new model for classifying dependency relations. For example, of the agreeing attribute in the phrase an interesting book, it is said: “The meaning of the noun in no way requires this property to be expressed; this property is not an essential element for understanding the meaning of the noun” (Holvoet, Judžentis 2003: 14). There seems to be no reason to deny this property in the case of a non-agreeing attribute.

6 “Genitives expressing matter, origin, price, or content are not required by the sense of their governing nouns” (Spevak 2014: 186).

DAIVA ŠVEKAUSKIENĖ. Valdymas ar šliejimas: kilmininku reiškiamas nederinamasis pažyminys ir prielinksninė konstrukcija

221

In the academic grammar, where it is stated that adjoining is a relation in which a dependent constituent joins a head without being required by it, and it is also asserted that in practice adjoining is often considered a relation in which “an invariable form joins the head constituent” (LKG 1976: 12), noun phrases with the genitive are regarded as instances of government, although examples of inflected forms that are adjoined are also given here. For instance, it is written that adjoining is most characteristic of adverbs, but it is then added: “In their function and relation to the head word, adverbs correspond to case forms with an adverbial meaning” (LKG 1976: 17). In the example provided, šią naktį lijo, the noun naktį is an inflected form adjoined to the verb as a temporal adjunct. Such an accusative with an adverbial meaning is essentially no different from a genitive functioning as a non-agreeing attribute: it is an inflected form that is not required by the head constituent of the phrase. Other cases expressing a locative adjunct, such as the locative and the instrumental of some words denoting location (e.g., patvoriais, paežere, paupiu), are likewise adjoined rather than governed.

3. PREPOSITIONAL CONSTRUCTIONS

In some sources, one can find claims that a case is also governed by a preposition. The Encyclopaedia of Lithuania states: “In Lithuanian, case forms are most often governed by verbs (e.g., matau namus, noriu duonos), prepositions (e.g., prie lango, per jūrą” (VLEₑ). A. Holvoet also discusses prepositional government: “<…> prepositional phrases are the clearest examples of case government” (Holvoet 2009: 115).

However, much earlier sources already stated: “Government by a preposition <…> is the joining of words to the words governing them by means of various prepositions” (Kalinauskas 1961: 33). The academic grammar of the Lithuanian language states: “Two or more words with independent lexical meanings, connected by a direct syntactic relation and forming part of a sentence, are called syntactic phrases. <…> A phrase can consist only of words with independent lexical meanings: verbs, nouns, adjectives, numerals, pronouns and adverbs” (LKG 1976: 9). A preposition is not an independent part of speech; it is merely a function word indicating relations between words in a sentence or phrase, and V. Labutis explains this very clearly: “In Lithuanian, <…> relations between words are indicated primarily by word forms (especially cases), then by function words (prepositions and conjunctions), and only lastly by word order” (Labutis 1998: 25). V. Labutis disagrees with the cla-

222

BENDRINĖ KALBA | 97

im that a preposition governs a case: “Thus, it is not appropriate to call the relation of a preposition, a non-independent word, to a case form government. If we assume that a preposition governs a case <...>, the concept of prepositional government becomes ambiguous: both the verb (or another word) governs a case with a preposition, and the preposition itself ‘governs’ that same case” (Labutis 1998: 35).

Thus, the Information System of the Lithuanian Language Grammar will follow the position long established in Lithuanian linguistics that a preposition does not govern a case. When a verb governs several cases, they are all equally required—dovanoti kam? ką? ‘give to whom? what?’ If a preposition is used with several cases (už – ko?, ką?; po – ko?, ką?, kuo?), in a sentence it always occurs with only one of those cases; which one is determined by the verb, not by the preposition itself.

A preposition–case construction, like a case form, can be either governed by or adjoined to the head constituent of a phrase: if it is required to complete the meaning of a word, it is governed; if it joins a word without being required by it, it is adjoined. Thus, in the phrase rašo laišką ‘writes a letter’, we would identify government, whereas rašo į Vilnių ‘writes to Vilnius’ involves adjoining. The examples of prepositional government given in VLEₑ are prepositional constructions that can be either governed or adjoined: prie lango – Jonas ilgai stovėjo prie lango ir mąstė ‘by the window—Jonas stood by the window for a long time and thought’ is adjoining, because stovėti ‘to stand’ does not require the prepositional construction prie + genitive, but there are words that do require it, for example, privažiuoti: automobilis privažiavo prie sankryžos ir sustojo. ‘to drive up to: the car drove up to the intersection and stopped.’ More examples can be found: the prepositional construction į + accusative is adjoined to the word siųsti (siųsti į Vilnių, į Lietuvą, į užsienį) ‘send (send to Vilnius, to Lithuania, abroad)’, but it is governed by the word panardinti (panardinti į vandenį, į alyvą) ‘immerse (immerse in water, in oil)’, and so on.

CONCLUSIONS

1. The genitive of a noun connected by a direct syntactic relation to another noun can be either governed by or adjoined to it. This depends on the meaning of the head word: nouns denoting measures, numerals and some other parts of speech require a complement, so the genitive is governed. A noun does not require a non-agreeing attribute, just as it does not require an agreeing attribute; therefore, a non-agreeing attribute expressed by the genitive is adjoined.

2. Some hold that a preposition governs a case. However, the position that only independent parts of speech can be linked by syntactic relations has long been established in Lithuanian syntax. A preposition is a function word that indicates in greater detail the relation between words with independent meaning.

DAIVA ŠVEIKAUSKIENĖ. Valdymas ar šliejimas: kilmininku reiškiamas nederinamasis pažyminys ir prielinksninė konstrukcija

223

REFERENCES AND SOURCES

Aliyeva Narmin 2014: A View on the Syntactical Relations. – American Journal of Linguistics 3(2), 41–45.

Ambrazas Vytautas 1986: Lietuvių kalbos sakinio sintaksinės ir semantinės struktūros vienetai. – Lietuvių kalbotyros klausimai 25, 4–44.

Babickienė Zofija, Venckutė Regina 2013: Kalbos mokslo pagrindai: vadovėlis, Vilnius: Mykolo Romerio universitetas.

Balkevičius Jonas 1963: Dabartinės lietuvių kalbos sintaksė, Vilnius: Valstybinė politinės ir mokslinės literatūros leidykla.

Biava Christina 2020: What is the Difference between Modifier and Complement? Prieiga internete: https://www.quora.com/What-is-the-difference-between-modifier-and-complement.

Bowers John 2001: Syntactic Relations. Prieiga internete: https://conf.ling.cornell.edu/bowers/SyntRelCompl.pdf.

Bučienė Laimutė 2014: Lietuvių kalbos sintaksė. Teorija ir praktika: mokomoji knyga, Vilnius: Lietuvos edukologijos universiteto leidykla.

DLKG 1997: Dabartinės lietuvių kalbos gramatika, Vilnius: Mokslo ir enciklopedijų leidybos institutas.

Grenda Česys 1964: Dėl veiksmažodinių junginių klasifikacijos. – Lietuvos TSR aukštųjų mokyklų mokslo darbai. Kalbotyra 10, 137–142.

Henisz-Dostert Božena, Macdonald Ross R., Zarechnak Michael 1979: Machine Translation. – Trends in Linguistics. Studies and Monographs 11, ed. W. Winterk, Hague, Paris, New York: Mouton Publishers.

Holvoет Axel 2009: Bendrosios sintaksės pagrindai, Vilnius: Asociacija „Academia Salensis“.

Holvoet Axel, Judžentis Artūras 2003: Sintaksinių ryšių tipai. – Lietuvių kalbos gramatikos darbai 1: Sintaksinių ryšių tyrimai, Vilnius: Lietuvių kalbos institutas.

Judžentis Artūras 2022: Prijungiamasis ryšys sovietų okupacijos laikų lietuvių sintaksės darbuose. – Bendrinė kalba 95, 133–161.

Kay Martin, Gawron Mark, Norvig Peter 1994: Verbmobil: A Translation System for Face-to-Face Dialog, Stanford: CSLI.

Kalinauskas Bronius 1961: Žodžių junginys kaip sintaksės objektas. – Tarybinė mokykla 5, 31–39.

Kalinauskas Bronius 1963: Žodžiai ir žodžių junginiai sakinyje. – Kai kurie gramatikos ir stilistikos klausimai: medžiaga mokytojui, Kaunas: Valstybinė pedagoginės literatūros leidykla.

Kalinauskas Bronius 1972: Lietuvių kalbos žodžių junginių sintaksė: mokymo priemonė, Vilnius: Valstybinis pedagoginis institutas.

Labutis Vitas 1998: Lietuvių kalbos sintaksė, Vilnius: Vilniaus universiteto leidykla.

224

BENDRINĖ KALBA | 97

LIGIS – Lithuanian Language Grammar Information System. Available online: https://ligis.lki.lt.

LKG 1976: Lietuvių kalbos gramatika 3, Vilnius: Mokslas.

P a u l a u s k i e n ė A l d o n a 2002: The separation of syntax from morphology in twentieth-century linguistics. – Kalbų studijos 3, 23–26.

PŽJ – Connecting word combinations. Available online: https://mokslai.lietuviuzodynas.lt/lietuviu-kalba/prijungiamieji-zodziu-junginiai.

S i r t a u t a s V y t a u t a s 1978. Pagrindinių sakinio dalių derinimas, Vilnius: Lietuvos TSR aukštojo ir specialiojo vidurinio mokslo ministerijos Leidybinė redakcinė taryba.

S l i ž i e n ė N i j o l ė 1994: Lietuvių kalbos veiksmažodžių junglumo žodynas 1, Vilnius: Mokslo ir enciklopedijų leidykla.

S l i ž i e n ė N i j o l ė 1998: Lietuvių kalbos veiksmažodžių junglumo žodynas 2(1), Vilnius: Mokslo ir enciklopedijų leidykla.

S l i ž i e n ė N i j o l ė 2004: Lietuvių kalbos veiksmažodžių junglumo žodynas 2(2), Vilnius: Lietuvių kalbos institutas.

S p e v a k O l g a 2014: Noun Valency in Latin. – Noun Valency, ed. O. Spevak, Amsterdam / Philadelphia: John Benjamins Publishing Company, 183–213.

Š v e i k a u s k i e n ė D a i v a 2013: Syntactic attraction of words. – Acta Linguistica Lithuanica 69, 130–153.

T e s n i è r e L u c i e n 1959: Éléments de syntaxe structurale, Paris: Klincksieck.

V a i č i u l y t ė - S e m ė n i e n ė L o r e t a 2001: Lithuanian adjective valency: doctoral dissertation, Vilnius: Vytautas Magnus University Press.

V a i č i u l y t ė - S e m ė n i e n ė L o r e t a 2023: „Prie raišūs“, „prieraišumas“ valent­ingumas dabartinėje lietuvių (rašto) kalboje. – Acta Linguistica Lithuanica 88, 181–205. Prieiga internete: https://www.lituanistika.lt/content/103312.

V a s k e l a i t ė R a m u n ė 2009: Sintaksinių ryšių modeliai lietuvių kalbos sintaksėje: daktaro disertacija, Kaunas: Vytauto Didžiojo universiteto leidykla.

V a s k e l a i t ė R a m u n ė 2003: Lietuvių kalbos daiktavardiniai junginiai sintaksinių ryšių požiūriu. – Acta Linguistica Lithuanica 48, 143–158.

VLEe – Universal Lithuanian Encyclopedia. Available online: https://www.vle.lt/straipsnis/valdymas-1/.

W i n o g r a d T e r r y 1983: Language as a Cognitive Process 1: Syntax, London: Addison-Wesley Publishing Company.

Ž i u g ž d a J u o z a s 1960: Lietuvių kalbos gramatika 2: Sintaksė, Kaunas: Valstybinė pedagoginės literatūros leidykla.

ŽS – Word relations. Available online: http://www.šaltiniai.info/index/details/606.

Received 2023-12-13 Accepted 2024-11-04

DAIVA ŠVEIKAUSKIENĖ. Valdymas ar šliejimas: kilmininku reiškiamas nederinamasis pažyminys ir prielinksninė konstrukcija

225

DAIVA ŠVEIKAUSKIENĖ

CONTROL OR ADJOINING: THE GENITIVE MEANS NON-AGREEING ATTRIBUTE AND PREPOSITIONAL CONSTRUCTION

Summary

When discussing syntactic relations in their most general terms, opinions usually do not differ much. Everyone admits that agreement is such a subordinating relation, when the dependent word acquires the forms of the head word (genus, number, case). Disagreements arise when defining control. There are different views on the control of the locative: from providing examples of this case in the control section to strictly stating that the locative cannot be controlled. In most publications, adjoining is defined as a relation when the dependent word joins the head word, the latter is not required and the adjoined word is not necessary for the meaning of the head word.

In works representing or oriented towards Western linguistics, two types of subordination are distinguished – complementation and modification. The criterion for their designation is the requirement or non-requirement of complementation. The requirement of complementation arises from the semantic insufficiency of the word. In the syntax of the Lithuanian language, three types of relations are distinguished: control, agreement and adjoining.

Due to the differences in opinions regarding subordinating relations, the relationship of a non-agreeing attribute with a head word is also assessed differently. It is customary to consider it as a control, but the noun expressed by the genitive in this case is not needed to supplement the meaning of another word (noun), therefore, when defining control as such a relationship, when a requirement is made, it should be considered only the relationship of certain nouns (denoting measure, capacity) with the genitive, and the relationship of an non-agreeing attribute with a head word should be assessed as adjoining.

The prepositional construction, like the cases, can be controlled and adjoined to the head word, for example, the prepositional construction to+accusative in the word combination dip in water is controlled, and it is adjoined in the word combination send to Vilnius.

A preposition cannot control a case, because syntactically related words are only content words. When a verb controls several cases, all of them are equally necessary and can all be in one sentence – dovanoti ką? (give what? – accusative) kam? (to whom? – dative). If a preposition is applied with several cases (už – behind+ genitive, for+accusative; po – after+genitive, what+accusative, under+instrumental),

226

BENDRINĖ KALBA | 97

it is always used with only one case, and which case is used determines the verb, not the preposition itself.

KEYWORDS: syntactic relationships, control, adjoining, genitive, non-agreeing attribute, prepositional constructions.

DAIVA ŠVEIKAUSKIENĖ Institute of the Lithuanian Language, 5 Petro Vileišio St., LT-10308 Vilnius [email protected]

DAIVA ŠVEIKAUSKIENĖ

GOVERNMENT OR ADJUNCTION: THE NON-AGREEING ATTRIBUTE EXPRESSED BY THE GENITIVE AND THE PREPOSITIONAL CONSTRUCTION

Summary

When syntactic relations are discussed in the broadest terms, opinions usually do not differ much. Everyone agrees that agreement is a type of subordinating relation in which the dependent word takes the forms of the head word (gender, number, case). Disagreement arises in defining government. Views differ on the government of the locative: in some works examples of this case are given in the section on government, while others express the firm view that the locative cannot be governed. In most publications, adjunction is defined as a relation in which the dependent constituent attaches to the head without being required by it, and the adjunct is not essential to the meaning of the head constituent.

Works representing or oriented toward Western linguistics distinguish two types of subordinating relation—complementation and modification. The criterion distinguishing them is whether a complement is required. The need for a complement arises from the semantic incompleteness of a word. Lithuanian syntax distinguishes three types of subordination: government, agreement, and adjunction.

Opinions differ regarding subordinating relations, and the relation between a non-agreeing attribute and the word it modifies is likewise assessed differently. It is usually regarded as government; however, a noun expressed in the genitive in this case is not needed to complete the meaning of another word (a noun). Thus, if government is explained as a relation that involves a requirement, only the relation of certain nouns (those denoting measure or capacity) to the genitive should be considered government, while the relation of a non-agreeing attribute to the word it modifies should be regarded as adjunction.

DAIVA ŠVEIKAUSKIENĖ. Valdymas ar šliejimas: kilmininku reiškiamas nederinamasis pažyminys ir prielinksninė konstrukcija

227

A prepositional construction, like a case form, may be governed by or adjoined to the head of a word combination; for example, the prepositional construction into + accusative in the word combination immerse into water is governed, whereas in the word combination send to Vilnius it is adjoined.

In Lithuanian syntax, a preposition is considered not to govern a case form; only words with independent meaning are linked by syntactic relations. When a verb governs several case forms, all of them are equally necessary and may all occur in the same sentence—give whom? what? If a preposition is used with several case forms (behind—what?; after—what?, what?, with what?), it always occurs in a sentence with only one of those case forms, and which one is determined by the verb, not by the preposition itself.

KEYWORDS: syntactic relations, government, adjunction, genitive, non-agreeing attribute, prepositional constructions.

DAIVA ŠVEIKAUSKIENĖ

Institute of the Lithuanian Language

5 Petro Vileišio St., LT-10308 Vilnius

[email protected]

228

BENDRINĖ KALBA | 97

Document details