Acta Linguistica Lithuanica

LX (2009), 1 – 15

Morphological Features of Spoken Lithuanian

Ineta Dabašinskienė

Vytauto Didžiojo universitetas

Research on spoken language is still a new and little investigated field in Lithuania. As the collection and analysis of natural spontaneous language data is a complicated task requiring special preparation and adequate methods of data collection, transcribing and coding, it requires considerable expenditure of time, financial means and personal efforts. This the reason why, until now, there has been no systematic research in Lithuania that would allow us to define the most important morphological, syntactic, semantic or pragmatic features of spoken Lithuanian.

The development of a grammatically coded corpus of spoken Lithuanian was the first step towards more extensive research in this direction. A comparison between spoken and written Lithuanian was the main aim of this study. The distribution of certain grammatical categories as well as the lexicon was studied in order to establish some tendencies in usage.

1. INTRODUCTORY NOTES OR WHO, WHERE, AND HOW SPEAKS…

Linguists often draw attention to spoken and written varieties of language, which differ in their syntactic, morphological, and lexical features. Spoken language, like written language, may vary in genre, which also determines linguistic characteristics; for example, television and radio broadcasts, lectures, interviews, etc., differ from colloquial speech¹, which is usually spontaneous, takes place at a particular time and in a particular space, and whose speakers have no opportunity to edit their speech in advance. The vocabulary of colloquial speech is less varied, and its syntax is fragmentary,

¹ Kazimieras Župerka discusses the everyday style, which he also calls colloquial, almost synonymously. The domains in which this style is used include speaking at home, on the street, and while traveling—that is, unofficial, non-professional communication characterized by simple interaction, an emotional-impact function, etc. (1997: 89).

1

Ineta Dabašinskienė

And the utterances are short and simple. It is easy to recognize by its use of nonstandard genders, dialectal forms, and slang, especially frequent repetitions, various forms of abbreviation, numerous fillers, and other linguistic units that are difficult to name and whose origin is hard to determine. According to Kazimieras Župerka (1997: 89–91), two opposing tendencies are at work here: linguistic shortening, associated with rapid speech, spontaneity, information load, and convenience, and linguistic excess, which indicates the excessive use of certain lexical elements.

Discussing textual variety and its markers, Andrius Utka argues that spontaneous speech, being highly expressive, has distinctive features that set it apart from other genres, and illustrates this with specific data: “The paradigm of spontaneous expressivity, which expresses the function of spontaneous expressivity, is the strongest of all, since it comprises as many as 32 significant features: 30 of them are positive (tai, jau, čia, ten, tas, ką, tą, tiek, taip, mes, labai, bet, dabar, aš, kaip, kur, nors, vis, tada, daug, nes, o, kai, kiek, to, buvo, gal, visi, jeigu, dar) and 2 are negative (average word length, iteration index)” (Utka 2004).

Despite the differences, which seem to suggest that colloquial speech is simpler, its complex structure should be noted. This complexity first becomes apparent when grammatical phenomena are identified. Clearly, there are difficulties in describing colloquial speech on the basis of theories (especially syntactic ones) developed according to the conventions of written language. Conversation or dialogic speech, depending on the situation, the participants in the conversation, their number, the topic, and other linguistic and non-linguistic features, has a structure of its own, different from that of written language, which is based on the linguistic flow of sentences.

In conversation, communication is important, and the latter depends on the speaking situation, which consists of the participants’ roles and relationships, as well as the place, time, and general communicative environment. The most important component of the speaking situation is its participants, whose relationships become apparent through interaction, thereby also revealing their social roles. Social roles, determined by changing situations, are realized through communication (Karaliūnas 1997: 88-89); it is therefore natural that many linguistic features are determined by the speaker’s situational social role. As an individual moves from one role to another, their speech and code also change: for example, the role of an institutional employee or director is mostly associated with formal, business style; the roles of colleague and customer are more characteristic of the ordinary, casual means of colloquial speech, while the social role of a friend is associated with the familiar, everyday conversational genre. Another important characteristic of the participants is the psychological aspect, that is,

2

Šnekamosios lietuvių kalbos morfologinės ypatybės

their attitudes, motives, emotions, likes and dislikes, and experiences (Versch ueren 1987). All these characteristics associated with the participants in a conversation determine not only its topic, which is related to lexical diversity, but also its pragmatics, which may reflect differences in morphology, syntax, and phonology. Thus, analyzing different conversations can yield different results, since these depend on the characteristics listed above.

2. RESEARCH MATERIAL AND METHODOLOGY

2.1. Corpus of Spoken Lithuanian² (ŠKT)

The development of corpora and their benefits are obvious to everyone. Today, we have large corpora of written Lithuanian, and more than one. Creating and compiling them is not as difficult as creating corpora of spoken language. The work involved is incomparable in terms of time and human and financial resources, which is why there are not many spoken-language corpora, and they are usually small. Spoken-language research is nothing new worldwide, especially for major languages. We have known about spoken English corpora for many years; corpora also exist for German, Croatian, Russian, Spanish, and even Estonian, but they vary in size—from a few thousand to several million words. Recently, a 10-year project in the Netherlands, involving substantial human and financial resources, was completed and announced the National Corpus of Spoken Dutch³, comprising 10 million words.

Research on spoken language in Lithuania is still a new and little-studied field. Since collecting and processing data on natural, spontaneous language is a complex task requiring special preparation, considerable time and effort, and financial resources, until now data collection has been limited to certain areas of spoken language (radio and television broadcasts—that is, conversations prepared/planned in advance rather than spontaneous ones). It is important that the idea of a spoken-language corpus, cultivated for many years, has been partially implemented

² Corpora of spoken and conversational language can be found online at www.vdu.lt/LT-courses/ --> SCIENCE

³ http://www.elis.rug.ac.be/cgn/ and http://lands.let.kun.nl/cgn/ehome.htm

3

INETA DABAŠINSKIENĖ

ted. The first studies began after financial support was received from LVMSF⁴, and a grammatically annotated corpus of spoken Lithuanian was compiled. A digital database of spoken language—that is, its transcribed, annotated, and audio forms—is the first step toward documenting and preserving the actual use of our native language for future generations. Although still small, this spoken-language corpus already makes it possible to observe and identify the main morphological, syntactic, semantic, and pragmatic features of spoken Lithuanian.

To ensure that the corpus is versatile and suitable for a wide range of analyses, it was decided to follow the principle of a balanced corpus. Thus, several criteria were considered in compiling the balanced corpus: the nature and structure of spoken language, the relationship between interlocutors, demographic indicators, and the social relationships between interlocutors. The corpus provides spoken-language data from Vilnius and Kaunas (and, as the work continues, from other regions of Lithuania⁵), together with audio recordings. The interlocutors’ gender, education, and age were taken into account.

2.2. Methodology

The study uses approximately 30 hours of recorded spoken language (natural spontaneous conversation), which has been transcribed and coded and comprises a total of about 50,000 morphologically annotated words. The recordings were made using a digital voice recorder capable of producing sufficiently high-quality audio recordings, easily transferring them to other digital media using the Windows Media Player program, and saving them in *mp3 format.

After the recordings had been made, the next stage of the work was their transcription and coding using a special program. CHILDES (Child Language Data Exchange System, MacWhinney 2000) is currently well known worldwide; the program was originally crea-

⁴ A project funded by LVMSF, “Developing a Corpus of Standard Spoken Lithuanian: Digitization and Grammatical Annotation” (VDU, 2006). The project was coordinated by VDU, with VU as a partner institution. The corpus comprises approximately 50,000 grammatically annotated words.

⁵ The LVMSF-funded project “Developing a Corpus of Spoken Lithuanian” continues the work already begun and collects and processes spoken Lithuanian data. In addition to VDU, VU, KU, ŠU, and LKI are participating in this project. The planned size of the corpus is approximately 250,000 grammatically annotated words.

4

Šnekamosios lietuvių kalbos morfologinės ypatybės

tually created to analyze child language data⁶, and has now also been successfully adapted for research on adult spoken language. One of the most important advantages of this program is that it enables the rapid and accurate automatic linguistic analysis of large volumes of text. The CHILDES system comprises two tools. The first is the CHAT transcription and coding format. The second is the CLAN analysis software package. Using the CLAN programs Combo, Kwal, and Freq, statistical analyses can be conducted on the frequency of the vocabulary and grammatical categories used, utterance length, and so on. By adapting the CHILDES program⁷ and its coding principles to Lithuanian, it will be possible in the future to conduct not only morphological but also syntactic analyses of spoken language, and to investigate discourse and pragmatic features of conversation.

2.3. Corpus Structure

The spoken-language corpus used in this study is currently relatively small, so it was decided to distinguish only the two most important social settings—the private and the public sphere. The private sphere is characterized by situations in which speakers have close, friendly, familiar social relationships. The most common social roles in private conversations are those of family member and friend. Communication in the public sphere is characterized by more formal relationships, and speakers’ social interaction most often takes place in various institutions—shops, hair salons, banks, schools, and universities. The most common social roles in public-sphere conversations are those of salesperson, customer, employee, client, teacher, and student. Since conversations in the private sphere most often take place among people who know one another well and communicate familiarly, they cover a variety of topics. The part of the corpus comprising these conversations is called “private conversations.” Communication in the public sphere most often takes place in various institutions; therefore, the part of the corpus comprising conversations in this domain is called “conversations in institutions” (see Table 1).

⁶ Data on children’s, child-directed, and, in part, adult language have been coded and analyzed using the CHILDES program for many years (Savickienė 2001, 2003, 2005; Balčiūnienė 2005, 2007).

⁷ Dr. Andrius Utka (VDU) adapted this program for the analysis of spoken Lithuanian.

5

Ineta Dabašinskienė

Table 1. Most frequent social roles of the participants in recorded conversations and locations of the conversations

Private conversations Conversations in institutions Participants’ social roles Conversation locations Participants’ social roles Conversation locations friend at home hairdresser at the hair salon husband, wife visiting seller at the bank son, daughter in the countryside doctor at the ticket office grandchild in a café patient in a shop grandfather, grandmother in a car manicurist at the beauty salon relative by phone client at the university acquaintance supervisor in the office employee at the pharmacy at the station

Many linguistic features depend not only on the conversational situation and the social roles of the participants, but also on factors such as the speaker’s gender and age (which are not taken into account here for the time being). Adults speak one way with other adults, people their own age or of a similar age, but differently with children and the older generation. When speaking with children, they modify and simplify their speech (Kamandulytė 2005, 2007, Savickienė 2006). When speaking with the older generation, various politeness modifications are characteristic, as is more frequent use of certain pronouns (Rosinas 1996: 41), a more respectful tone, clearer speech, and a particular vocabulary. Speakers’ gender is also very important for the sociolinguistic and pragmatic analysis of conversation. The linguistic features of speakers with different levels of education also differ—in terms of correctness, expressiveness, verbal fluency, and lexical diversity. The distribution of participants by gender and age is very varied and is described in Table 2.

In order to create a small but balanced corpus of standard spoken Lithuanian, the structure of conversations was also taken into account. Simas Karaliūnas (1997: 105) distinguishes the following structural types of communicative acts—dialogue, monologue, conversation (polylogue), discussion, lecture, presentation, correspondence, theatrical performance, and so on. Since the subject of this study is spontaneous spoken Lithuanian, only types characterized by spontaneity were collected: dialogues and polylogues. The nature of spoken language is related to conversational structure: a spontaneous conversation takes place between two or more participants; thus, spontaneous speech is associated with the distinction between dialogues and polylogues, whereas prepared—

6

Šnekamosios lietuvių kalbos morfologinės ypatybės

public speech is delivered by one person, and is therefore associated with monologue. Of course, a monologue often turns into a dialogue or polylogue when the speaker is asked questions or engaged in discussion, so no strict boundary should be drawn between spontaneous dialogue and polylogue, on the one hand, and public monologue, on the other. Conversation topics are not described because they are highly varied and often change many times even within a single conversation. The recorded texts are distributed by conversational structure as follows: 25 dialogues and 28 polylogues.

Table 2. Distribution of recorded conversations by participants’ age and gender

Age group

Number and gender of participants

children (1–10 years)

2 boys, 4 girls

adolescents (10–18 years)

2 boys, 3 girls

young adults (18–30 years)

12 men, 19 women

middle-aged adults (30–60 years)

8 men, 17 women

older (>60 years old)

1 man, 1 woman

Total:

25 men, 44 women

3. MORPHOLOGICAL FEATURES OF SPOKEN LANGUAGE

Data from the Corpus of Contemporary Written Lithuanian (RKT) show that the most frequent Lithuanian word forms are: and, Lithuania’s, Republic’s, because of, Seimas, or, either, from, and, with. The most frequent noun form is — Lietuvos, the adjective form — aukščiausiosios, the verb form — buvo, the most frequent pronoun — jis (www. donelaitis.vdu.lt). It is clear that this vocabulary is related to a specific source: the written-language corpus consists mainly of journalistic texts collected from newspapers and magazines.

As might be expected, the study data showed that the use of vocabulary and morphological categories differs in the spoken-language corpus. The most frequently used noun forms in the spoken-language corpus were found to be — time, day, time, place; the most frequent adjective forms — good (masculine), good (feminine), beautiful; the most frequent verb forms — I don’t know, I know, is, wants, to be, isn’t, to say; the most frequent pronouns — I, you, we, that; adverbs — there, here; and other words — well, and, not, but, that, what, it, or. These features show that spoken language is characterized by short words and shortened word forms.

7

INETA DABAŠINSKIENĖ

It is important to mention that when discussing vocabulary in general or the semantic groups of individual parts of speech, one must bear in mind that associations between meaning and form can often be identified (for case meanings, see Savickienė 2003, 2005). The frequent use of the nominative case indicates that nouns denoting people are used more often in ŠKT than in RKT. The meaning of first-person verbs (mental processes) differs from the meaning of verbs used in the third person (active action). Adherents of the usage-based approach usage-based approach and frequency-effect frequency effect theoretical approach to grammar, such as Michael Tomasello (2003), Joan Bybee (1985, 1995), and other scholars (Bybee & Hopper 2001), argue that the use of more or less frequent forms depends on the meaning of a particular word and, conversely, that certain words may be used more frequently only in certain forms. The influence of frequency can lead to changes in word structure, e.g., shortening and contraction. So far, few studies in Lithuania have examined in detail the influence of frequency on form and meaning.

Based on preliminary spoken-language corpus data, the main trends in the use of grammatical categories in contemporary spoken Lithuanian can be identified. This article sought to present an account of the morphology of spoken Lithuanian only, rather than of spoken Lithuanian in general; therefore, it examined only ŠKT, not texts from individual spoken-language genres. It is understandable that the use of grammatical forms and lexical units may differ in other spoken-language genres. However, as Vida Žilinskienė, who has long studied the use of grammatical forms in journalistic, technical, scholarly, and literary styles, argues, the differences are minor. She found that literary-style texts differ most from non-literary texts (Žilinskienė 2005). It may be assumed, of course, that spoken-language data of other kinds will not reveal major differences; however, that remains a matter for further research.

The data presented by Vida Žilinskienė (2001; 2002a; 2002b; 2005), who has long studied the distribution of morphological categories across different styles of written Lithuanian, as well as by Erika Rimkutė (2006), show that the findings of the two researchers are very similar despite the different texts and styles. For this study, we chose to compare the data from Erika Rimkutė’s morphologically annotated corpus as a source of written Lithuanian.

First, we wanted to examine how parts of speech are distributed in spoken language. The data in Table 3 show obvious differences between the ŠKT and RKT data.

8

Šnekamosios lietuvių kalbos morfologinės ypatybės

Table 3. Distribution of parts of speech

Part of speech

ŠKT

RKT

Noun

16,2

39,37

Adjective

2,8

7,33

Pronoun

16,9

8,7

Numeral

1,9

0,96

Verb

22,8

20,5

Adverb

10,2

6,72

Preposition

4,2

4,65

Particle

12

1,97

Conjunction

8,2

7,62

Ideophone, interjection

4,8

0, 2

Total:

100%

100%

First, it is evident that the use of the main parts of speech differs in frequency: the most frequent part of speech in ŠKT is the verb, whereas in RKT it is the noun, although the difference in verb use is not very large (22.8% and 20.5%). Nouns dominate in written language, while verbs dominate in spoken language. This result is not unexpected, since spoken language is much more expressive and lively, qualities fostered by more frequent use of verbs. Pronouns are used slightly more often than nouns in ŠKT; pronouns are also in third place in terms of frequency in RKT, but the frequency of this part of speech differs twofold, cf. 16.9% and 8.7%. Unlike RKT, ŠKT uses particles, adverbs, interjections, and ideophones much more often. These parts of speech are also associated with the expressiveness, spontaneity, and emotionality of the discourse being produced. As Audrius Utka states, “the frequency distribution of very frequent words or their forms in texts is neither chaotic nor random. Being the most frequent structural units of a text, frequent words are directly related to the functions of the text and, together with other formal features of the text, are therefore significant indicators of text functions” (2005: 49). Numerals, conjunctions, and prepositions are used with similar frequency. Very

9

Ineta Dabašinskienė

The use of adjectives differs: they are used much less frequently in the SPOC than in the WPC. Other studies of spoken language analyzing conversations between children and adults also show that adjectives are used particularly rarely (Kamandulytė 2005).

It would be useful to discuss and compare the distribution of the grammatical categories of nouns and verbs, the most frequent parts of speech. Adjectives will also be discussed because they use grammatical categories similar to those of nouns, although this part of speech is not used frequently.

Case. Grammatical cases are used more frequently than semantic cases in both the SPOC and the WPC (see Table 4). Thus, cases fall into two groups according to frequency of use: a core group (nominative, genitive, and accusative) and a peripheral group (dative, instrumental, locative, and vocative). More obvious differences emerge when the frequency of use of each case is examined. The nominative is the most frequent noun case in the SPOC, whereas the genitive is the most frequent in the WPC. This difference is particularly pronounced, as the frequency of the singular genitive differs by almost a factor of two (cf. 22.9% and 40.2%). The frequent use of the genitive in the WPC is due to its descriptive genitive function, which it performs in numerous journalistic sources. The accusative is used much more frequently in spoken than in written language (cf. 26% and 15.8%). Thus, the order of the main singular cases by frequency in the SPOC is nominative, accusative, genitive, while in the WPC it is genitive, nominative, accusative. Earlier studies of spoken language recorded frequent use of the nominative in the speech of both children and adults (see Savickienė 2003, 2005). As expected, the use of the singular locative and instrumental is similar in the SPOC and the WPC, but the vocative and dative differ: the vocative is much more frequent in spoken language (cf. 3.3% and 0.6%), while the dative is more frequent in written language (2.6% and 4%). In the plural paradigm, we see only a somewhat more frequent use of the genitive in the SPOC, while the distribution of the other cases is similar to that in the singular paradigm.

Table 4. Distribution of noun and adjective cases (%).

Case SPOC Noun Sg. SPOC Noun Pl. SPOC Adj. Sg. SPOC Adj. Pl. WPC Noun Sg. WPC Noun Pl. WPC Adj. Sg. WPC Adj. Pl. Nom. 31.6 30 62.5 61.7 27.5 23.7 40 34.6 Gen. 22.9 33.5 15.3 17.7 40.2 42.4 26.7 30.3

10

Šnekamosios lietuvių kalbos morfologinės ypatybės

Continuation of Table 4

Case SPOC Noun Sg. SPOC Noun Pl. SPOC Adj. Sg. SPOC Adj. Pl. WPC Noun Sg. WPC Noun Pl. WPC Adj. Sg. WPC Adj. Pl. Dat. 2.6 2.3 1.5 1.3 4 4.3 3.3 3.9 Acc. 26 22.8 14 10.7 15.8 18.2 17.7 19.1 Instr. 5 9 4.4 7.7 5.8 7.3 7.8 8.5 Loc. 8.6 2.1 2 0.6 6 3.8 4.1 3.6 Voc. 3.3 0.3 0.3 0.3 0.6 0.3 0.2 0.08

Statistical data for Slavic languages are similar to the results for spoken Lithuanian: Russian: Nom. — 32.6%, Acc. — 25.3%, Gen. — 22%, Dat. — 4.1%, Loc. — 10.1%, Instr. — 5% (Zemskaja 1979: 74); Polish: Nom. — 34.2%, Acc. — 29.8%, Gen. — 19.2%, Dat. — 4.8%, Instr. — 4.4%, Loc. — 7.6% (Laskowski 1989: 212).

In the adjective paradigm, the lists of case frequencies in both the SPOC and the WPC are similar: the nominative is used most often, followed by the genitive, accusative, instrumental, locative and dative, and vocative. The greatest difference is evident in the use of the nominative and genitive: in the SPOC, the adjective nominative is used very frequently, reaching as much as 62.5%, whereas in the WPC it accounts for only about 40%; the genitive is less frequent in spoken than in written language: 15.3% and 26.7%. Thus, both noun and adjective paradigms show that the most frequent case in the SPOC is the nominative. It should be noted that in the SPOC there is no clear difference between the use of the noun and adjective accusative and genitive, since their singular and plural frequencies are similar. In the adjective case system, the genitive ranks second in frequency, whereas in the noun paradigm the accusative is more frequent in the singular, and the genitive in the plural.

Gender and number. The results showed that masculine gender, as the unmarked member, dominates in both corpora: in the SPOC, masculine nouns account for 55.9% and feminine nouns for 44.1%; masculine adjectives account for 55.6% and feminine adjectives for 44.4%. In the WPC, masculine nouns account for 60.4% and feminine nouns for 39.6%; masculine adjectives account for 54.4% and feminine adjectives for 41.5%. The singular is used more frequently than the plural.

11

INETA DABAŠINSKIENĖ

Verb forms. The verb is the most frequently used part of speech in the ŠKT. A comparison of finite and non-finite verb forms shows that finite forms make up more than 80% in the ŠKT, but about 50% in the RKT. Unlike in the RKT, participles, adverbial participles and half-participles are used much less frequently in the ŠKT.

Table 5. Distribution of verb forms

Forms | ŠKT | RKT; Finite forms | 82,1 | 50,7; Participles | 5,4 | 27,7; Adverbial participles | 0,4 | 3,2; Half-participles | 0,2 | 1,8; Infinitive | 11,9 | 16,6; Total | 100% | 100%

The distribution of tense usage was similar in both corpora: the present tense is used most frequently, accounting for half of all tense forms, followed by the simple past, the future, and the very rarely used habitual past. Future-tense forms occurred almost twice as often in the ŠKT as in the RKT, cf. 12,9% and 7,9%.

Table 6. Distribution of verb tenses

Tenses | ŠKT | RKT; Present | 55,6 | 52,5; Simple past | 30,9 | 37,8; Habitual past | 0,6 | 1,8; Future | 12,9 | 7,9; Total: | 100% | 100%

Among finite verb forms, the third person predominates. In the RKT, the third person dominates in both singular and plural forms. The ŠKT presents different data: the first and third persons singular are used almost as frequently

12

Šnekamosios lietuvių kalbos morfologinės ypatybės

as each other, and the second person is also not marginal in usage. The similar use of the first, second and third persons indicates that dialogic speech predominates.

Table 7. Distribution of verb persons

Persons

ŠKT

RKT

Sing. 1st person

28,4

7,5

Sing. 2nd person

21,2

4.6

Sing. 3rd person

29,2

53,4

Plur. 1st person

6

5,9

Pl. 2nd person

4,4

2,5

Pl. 3rd person

10,8

26,1

Total:

100%

100%

Verbs of the first conjugation predominate in Lithuanian, so it is no coincidence that the ŠKT data show that verbs of this conjugation account for more than 61% of usage, those of the second conjugation for 17%, and those of the third for 22%.

4. CONCLUSIONS

Based on preliminary data from the spoken-language corpus, the main trends in the use of grammatical categories in contemporary spoken Lithuanian have been identified. Having discussed the use of grammatical categories of the main parts of speech in the ŠKT, we can conclude that these data differ markedly from the RKT data. First, the very frequent use of verbs, pronouns, particles, and other function words should be noted. The very rare use of adjectives could serve as a suitable criterion for identifying spoken language. Other criteria are also important: the predominance of the nominative in singular and plural noun and adjective paradigms; the infrequent use of non-finite forms (especially participles); and the similar usage rates of the first and third persons of the verb.

These results are a first attempt to provide statistical support for differences between spoken and written language that have often been recorded only through observation. It is hoped

13

INETA DABAŠINSKIENĖ

that various studies of lexicon, morphology, and syntax based on spoken-language data will increase in number in the future.

LITERATŪRA

BALČIŪNIENĖ, I. 2005: Parodomųjų įvardžių įsisavinimas. Lituanistica (4), 45–54.

BALČIŪNIENĖ, I. 2007: Parental reactions to the grammatically correct and incorrect utterances of the children. Estonian papers in Applied Linguistics 3, 47–56.

BYBEE, J. 1985: Morphology: A Study of the Relation Between Meaning and Form. Amsterdam: John Benjamins.

BYBEE, J. 1995: Regular Morphology and the lexicon. Language and cognitive processes 10 (5), 425–455.

BYBEE, J. & P. HOPPER, eds., 2001: Frequency and the emergence of the linguistic structure. Amsterdam: John Benjamins.

KAMANDULYTĖ, L. 2005: Vaikiškosios kalbos registras. Gimtoji kalba 7, 12–16.

KAMANDULYTĖ, L. 2007: Morphological Modifications in Lithuanian child directed speech. Estonian Papers in Applied Linguistics 3, 155–167.

KARALIŪNAS, S. 1997: Kalba ir visuomenė. Psichosociologiniai ir komunikaciniai kalbos vartojimo bruožai. Vilnius.

LASKOWSKI, R. 1989: Markedness and the category of case in Polish. In: O. M. Tomić, ed., Markedness in synchrony and diachrony. Berlin etc: Mouton de Gruyter, 207–226.

MACWHINNEY, B. 2000: The CHILDES project: tools for analyzing talk: Transcription, format and programs. Mahwah, NJ: Lawrence Erlbaum Associates.

RIMKUTĖ, E. 2006: Morfologinio daugiareikšmiškumo ribojimas kompiuteriniame tekstyne. Daktaro disertacija. Kaunas: VDU.

ROSINAS, A. 1996: Lietuvių bendrinės kalbos įvardžiai. Vilnius: Mokslo ir enciklopedijų leidykla.

SAVICKIENĖ, I. 2001: The role of diminutives in Lithuanian child language acquisition. Linguistica Baltica 9, 109–118.

14

Šnekamosios lietuvių kalbos morfologinės ypatybės

Savickiene, I. 2003: The Acquisition of Lithuanian Noun Morphology. Wien: Verlag der Österreichischen Akademie der Wissenschaften.

Savickiene, I. 2005: Linkšnių vartojimo dažnumas ir daiktavardžių reikšmė. Acta Linguistica Lithuanica 52, 59–65.

Savickiene, I. 2006: Komunikacinė pragmatika ir kalbėjimo situacijos tikslas: deminutyvų vartojimo atvejis. Kalbos kultūra 79, 258–265.

Tomasello, M. 2003: Constructing a Language. A Usage Based Theory of Language Acquisition. Cambridge, MA: Harvard University Press.

Utka, A. 2004: Statistinis tekstų funkcijų nustatymas. Daktaro disertacija. Kaunas: VDU.

Utka, A. 2005: Labai dažnos žodžių formos lietuvių kalboje. Lituanistica 1(61), 48–55.

Verschu(e)ren, J. 1987: Pragmatics as a theory of linguistic adaptation. (IPrA Working Document 1).

Zemskaja, E. A. 1979: Russkaja razgovornaja reč’. Moskva: Nauka.

Žilinskiene, V. 2001: Lietuvių ir latvių kalbų publicistikos leksikos ir morfologijos pagrindinės statistinės charakteristikos. Lituanistica 3 (47), 69–79.

Žilinskiene, V. 2002a: Žodžių formų vartojimas lietuvių kalbos dalykinio ir publicistinio stilių duomenimis. Lituanistica 1 (49), 106–116.

Žilinskiene, V. 2002b: Gramatinių formų vartojimas lietuvių kalbos moksliniame stiliuje. Acta Linguistica Lithuanica 46, 173–183.

Žilinskiene, V. 2005: Vardažodžių, įvardžių ir jų gramatinių formų vartojimas lietuvių kalbos stiliuose. Lituanistica 4 (64), 28–44.

Župerka, K. 1997: Stilistika. Šiauliai: ŠU.

Ineta Dabašinskienė

Vytauto Didžiojo universitetas

Regionistikos katedra

K. Donelaičio g. 52-208, LT-44244 Kaunas

[email protected]

15