Null subjects in old Germanic languages

Размещено на http://www.allbest.ru/

Null subjects in old Germanic languages

H. Zinchenko

Kyiv National Linguistic University

Abstract

Key words: Null subject, subjecthood, focal prominence, Germanic languages.

Зінченко Г.Є.

Ключові слова: нульовий суб'єкт, суб'єктність, фокус уваги, германські мови.

Зинченко А.Е.

В статье рассматриваются подходы к определению универсальных характеристик субъектности в когнитивной, генеративной и функциональной парадигмах. В зависимости от типа нулевых субъектов различают три типа языков. Отмечается, что нулевые субъекты функционируют в древнегерманских языках и некоторых диалектах современных германских языков. Проанализирована синтаксическая дистрибуция древнегерманских нулевых субъектов, их обусловленность согласованием с глагольной формой в предложении, частотность употребления в зависимости от формы лица глагола. languages subjectness cognitive

Ключевые слова: нулевой субъект, субъектность, фокус внимания, германские языки.

Introduction

Grammatical subjecthood has been described in a number of linguistic schools and has been defined from different perspectives. The discussion of this type of grammatical relations is marked by the diversity of viewpoints and leads to a certain degree of theoretical disagreement as for its nature and status. '1 he questions arise concerning the conclusive ways of characterizing the subjects and the possibility of recognizing them in every language, i. e. the universality of the subject [7; 8; 11].

Theoretical Background

In functionally oriented approaches, subject is linked to semantic functions in terms of cases, prototypical semantic role (e.g. `agent' in active clause), or hierarchies of functions and roles [11, 7]. A number of characteristic properties are typically applied to identify subjects in different languages. Triggering verbal agreement, binding anaphors, and the ability to undergo raising to subject and to object are among the most frequently noted canonical subject properties [1,6; 11, 1-2]. However, grammatical behaviours of subjects differ from one language to the next, and these specific characterizations seem not to allow a unifying definition of the sentence constituent. In search of universal schematic definitions, cognitive grammar looks to the focusing of attention as one of the basic cognitive abilities. In view of varied language strategies, preferences, and conventions, different clausal participants may appear in focus of attention, that is, become prominent. R. Langacker aims at a more abstract and universal definition of subject and highlights the interplay between semantic roles and focal prominence as the aspects of conceptual organization. Semantic roles are pertaining to conceptual content; they are inherent in the structure of the occurrence, while focal prominence is a matter of construal and resides in the directing of attention. Typological language variation demonstrates different default choices of prominent elements (the two major options being the semantic roles of agent and patient) and the level of consistency in correlation between a particular semantic role and the focal prominence [7, 365-366]. The participants of a profiled relationship are conferred with different degrees of focal prominence. In particular, primary and secondary prominence degrees are schematic characterizations of subject and object respectively. Hence subject is defined as a nominal expression that specifies the primary focal participant, or the trajector of a profiled relationship, while other constituents are secondary figures, e.g. the object that specifies the landmark. There may be different candidates for primary focal prominence. For instance, imposing the tra- jector status on the agent or patient results in active or passive structures, as in Floyd broke the glass vs. The glass was broken by Floyd [7, 367-369].

Following Langacker's presentation of the conceptual structure of the nominal group and the clause as the three-layered construct consisting of type specification, instantiation, and grounding, M. Taverniers interprets subject as the Instantiator (a clausal element that instantiates a certain semantic role) and claims that it is the "primary syntagm-forming element for realizing processual meanings" [11, 10-19].

Methods

The idea of subject prominence is reflected in various versions of generative grammar. Initially, subject was presented as the constituent that takes the first level position below the top of a tree sentence diagram [S NP (= Subject) VP]; in later versions of generative theory, it is analyzed as the external argument. The generation of subject remains at issue, which leads to the diverse views of its position: as specifier of IP (inflection), VP, Fin (finiteness), TP (tense), or EventP [4; 8; 11].

Canonically, subject is represented by a nominal group in the nominative case. There are languages, however, that can leave the subject of a sentence unexpressed (= null). The main types of null subject languages are distinguished according to the kind of unexpressed subject they admit and structural configurations that allow null subjects [3, 4-7]. According to H. A. Sigurdsson, there are three types of languages: the Romance pro drop type (1), the Germanic topic drop type (2), and the Chinese discourse (= radical) drop) type (3) [10, 269].

(1) It Parlo/Parli islandese.

speak 1SG/2SG Icelandic

`I/You speak Icelandic.'

(2) Sw Kommer tillbaks imorgon.

come 0 -AGR back tomorrow

`[I/We/She, etc.] will be back tomorrow.'

(3) Chi Kanjian ta le.

see 0 -AGR he PERF 0 -AGR

`[He/She, etc.] saw him.'

Results and Discussion

If a full, referential subject can remain unexpressed, the language is referred to as a canonical / full null subject language or a full pro-drop one. Greek, all Romance languages excluding French, Turkish, Arabic (in main clauses) are considered to be canonical null subject languages. The category of pro denotes an empty pronoun that fills in the canonical subject position [3, 2-3]. In this sense, the parameter is regarded as a morphological/lexical one, rather than a purely syntactic one. For instance, in Italian, virtually any clause can be uttered leaving the subject unexpressed, though sentences with overt subjects are also perfectly grammatical. Here, the agreement on the verb is analyzed as an incorporated pronoun, with first and second person matching the speaker and addressee features and third person matching an aboutness-shift topic [3, 1; 10, 273; 12, 272]. Example (4) demonstrates the difference between a sentence in Italian and the corresponding sentence in English, which by contrast does not admit a null subject.

(4) It (Voi) state leggendo un libro.

Eng *(You PL) are reading a book.

Languages that allow unexpressed subjects, as well as objects, and do not have any verbal inflection are defined as radical pro-drop or discourse prodrop languages. Here belong many Asian languages: Chinese, Japanese, Korean, Thai, Vietnamese etc. [2; 3, 3; 10, 269-270; 12, 272]. The following sentences in Vietnamese (5) demonstrate that arguments can be omitted quite freely without any agreement marking on the verb or on other verb related functional elements [2]. The verb is a bare stem, lacking person and number inflection. Both subject and object can remain unexpressed if they can be recovered through antecedents accessible in the discourse context.

(5) Viet Mary thich Tom. Va 0 cung thich Peter.

Mary like Tom And 0 also like Peter

Maryi likes Tom. Shei also likes Peter.'

Viet Mary thich Tom. Nhting Peter khong thich 0

Mary like Tom But Peter NEG like 0

`Mary likes Tomi. But Peter doesn't like himi.'

According to H. A. Sigurdsson, Romance (prodrop) null subjects differ from the Germanic (topic drop) and the Chinese (discourse = radical drop) types in being conditioned by verb agreement. Discourse drop languages have no clause-internal restrictions. Germanic null subjects, in turn, structurally differ from the other types in being restricted to clauses with an empty Spec, C. Germanic referential null subjects must thus have access to Spec, C [10, 268-270]. Sentence (6) in Swedish exemplifies Empty Spec, C position, but no agreement.

(6) Sw Kommer tillbaks imorgon.

come 0 -AGR back tomorrow

`[I/We/She, etc.] will be back tomorrow.'

In partial null subject languages, the number of contexts, allowable for unexpressed subjects, is limited. The parameter is restricted to specific structures or feature composition of subject pronouns. Null subjects are optional in some contexts where they are obligatory in full (= consistent) null-subject languages and excluded in non-null-subject languages, and excluded in some contexts where they are allowed in consiste nt null-subject languages. [3, 4-5; 5, 59-60]. For example, in Finnish, which is a partial null subject language, not all referential subjects can be null: first and second person subjects can be omitted, but third person cannot, compare (7) and (8):

(7) Finn (Sina)puhut englantia

`You speak 2 SG English.'

(8) Finn *(He) puhuvat englantia

`They speak 3 PL English.'

The difference is also made between thematic and non-thematic subjects, i.e. argumental, fully referential DP subjects and expletive subjects. Expletive null subject languages admit an unexpressed expletive subject, but do not allow a null referential subject. Partial and expletive null subject languages are different in that in the former subject omission is determined by syntactic conditions, while in the latter it is determined by the nature (referential or expletive) of the subject [3, 7]. In the following examples from Brazilian Portuguese (9) and Finnish (10), the weather-predicates do not license a theta-marked subject and have no overt expletive subject [5, 62].

(9) BP Esta chovendo.

is raining

`It's raining.'

(10) Finn Ulkona sataa. outside rains `It's raining outside.'

What concerns Germanic languages, null subjects are attested in different periods of their development. However, H. Rosenkvist states that unexpressed subjects are not admitted by standard modern languages and can be found in Old Germanic languages: Old English (11), Old High German (12), Old Icelandic (13), and Old Swedish (14) and a number of Modern Germanic vernaculars: Bavarian (15), Schwabian, Zurich German (16), Frisian, Yiddish, and Ovdalian [9, 152].

(11) OE Sona fat gesawon (Beowulf)

soon that saw

`Soon they saw that'.

(12) OHG Druthin ist auh

Lord is also

`He is also the Lord'.

(13) Olce fa skar Rognvaldr jarl [har hans],

en abr hafbi verit Uskorit

then cut R. jarl hair his but before had been uncut `Then R. cut his hair, but it had been uncut before'.

(14) Osw far gierfi kirchiu afra

there built church other `There he built an other church'.

(15) Bav ...obst noch Minga kummst

if 2 SG to Munich come 2 SG

`whether you come to Munich'

(16) ZG Wann nach ZUri chunnsch, muesch mi bsueche.

when to Ztirich come 2 SG must 2 SG me visit `When you come to Zurich, you must visit me'.

The discussion of null subjects primarily focuses on their syntactic distribution, relation to verb agreement, person reference, and frequency as compared to overt subjects. In this article, we will further outline the properties of Old Germanic languages.

Although Old Germanic languages are marked for a certain degree of syntactical similarity, they demonstrate unequal rates of null subjects, which is apparently explained by the different age of the languages and hence, the different stage of the overall Germanic diachronic development reflected in them [9, 153]. It has been noted that certain aspects of the evidence from early Icelandic contrast with findings from other early Germanic languages. Icelandic is marked for its longevity of the null subject property in a cross-Germanic perspective, unlike other Germanic languages, including English, German and Swedish, that lost the null subject property at much earlier stages. Moreover, null subjects show higher frequency in early Icelandic than in Old English and Old Swedish, where occurrence of null subjects is restricted [6, 37-38].

Old Germanic verb forms feature a rather strong agreement morphology. Being so distinct, they may provide supplementary information about the subject (subject pronoun), which then may become redundant and remain unexpressed. This idea leads to the assumption that subject omission is related to the richness of verbal inflection. However, a number of scholars claim to provide the evidence for the opposite [6, 39; 9, 154-160; 13]. In Old English, for example, verb forms that help unambiguously identify an omitted subject are singular ones, while plural forms do not have person distinction. Irrespective of this fact, both singular and plural subjects appear as null constituents, see (11) and (17), (18) [12, 279-280; 13, 161].

(17) OE ic swcgepepwt to disse niht wrpon hona

crwd priowa 0 me onswcest. (Rushworth, Matthew 26.34)

I say you that on this night before cock crows thrice 0 (pro) me deny 2 SG `I say to you that this night before the cock crows you will deny me three times.'

(18) OE pa ladde mon ford sumne blinde mon.

then led man NOM forth some ACC

blind ACC man ACC Was 0 wrest lwded to Bretta biscopum (Bede) was first led to Britons GEN bishops DAT `Then someone led forth a blind man.

He was first led to the priests of the Britons'.

Person becomes a more distinguishing factor, as third-person subjective (11), (18) arguments were dropped more frequently than those of the first and the second person (17) [9, 56; 12, 279; 14, 1]. Concerning the person reference, Old High German, Old Icelandic, and Old Swedish are similar to Old English: third-person null subjects are more numerous, for example (19), (20).

(19) OHG 0 steih tho in skifilin (Tatian 193.1)

stepped 3 SG then into boat `He then stepped into the boat'.

(20) Olce ok kom hann pangat ok var Hoskuldr uti,

er 0 reid і tun.

and came he there and was H. outdoors when rode into field `And he came there, and H. was outdoors when he rode into the field'.

Another argument in favour of that unexpressed subjects were not dependent on or facilitated by distinct verb agreement is that they were largely lost in Old High German period, although the language distinguished six verb forms marked for person and number and featured no substantial weakening of inflectional endings in its diachronic development. Likewise, Icelandic referential null subjects disappeared during the 18th and 19th centuries without any related changes in the verb agreement paradigm. Furthermore, in Old Swedish, singular null subjects were more common than plural, although there was only one verb form for singular [9].

There is less consistency in distributional features of unexpressed subjects in the discussed languages. In Old English and Old Swedish, null variants are more typical for main clauses, as in (11), (14), (18), than for subordinate ones, see (17), (21).

(21) OE godfremmendra swylcum gifepe bid pwt good-doers GEN such DAT given is that pone hilderws hwl gediged (Beowulf) the ACC battle-charge ACC hale endure `To such performers of noble deeds it will be granted that they survive the assault unharmed'.

Old High German embedded clauses with null subjects are registered with main clause word order, and clauses with the finite verb in final position license an overt subject. This fact allows generalization that Old High German null subjects are restricted to main clauses. Conversely, in Old Icelandic sentences display slightly rarer occurrences of dropped subjects in main or conjunct clauses (13) as compared to subordinate ones (20) [6, 4-17, 38; 9, 155-156; 12, 276; 13, 163-173].

Conclusions

The analysis of null subject characteristics in Old Germanic languages suggests that they show some similarities. Overall, null subjects are more common in main clause contexts, though they are registered in subordinate clauses as well. Rich verb paradigms do not seem to be the crucial factor for licensing null subjects. The distribution of unexpressed subjects is more determined by the person reference, as third-person omitted arguments are most frequent. Finally, gradual loss of null subjects is not related to the diachronic decrease of the number of distinct verb forms because it is also observed in languages that preserved verb inflections.

References

1. Citko, B., et al. (2018). If You cannot Agree, Move on! On labels and non-nominative subjects. Ubiquity Press, Glossa: a journal of general linguistics, Vol. 3(1): 28, pp. 1-46. DOI: https://doi.org/10.5334/gjgl.399

2. Cognola F., Casalicchio, J. (2018). On the Null-Subject Phenomenon Oxford: Oxford University Press, Null Subjects in Generative Grammar: A Synchronic and Diachronic Perspective [ed. by F. Cognola, J. Casalicchio], pp. 1-28.

3. D'Alessandro, R. (2014). The Null Subject Parameter Where are we and where are we headed? 27 p. First draft; published as: D'Alessandro, R. (2015). Null Subject. Contemporary linguistic parameters [ed. by A. Fabregas, J. Mateu, M. Putnam], London: Bloomsbury Press, pp. 201-226. https://ling.auf.net/lingbuzz/002159

4. Harley, H. B. (1995). Subjects, Events and Licensing. PhD dissertation, MIT: Department of Linguistics and Philosophy, 239 p. https://dingo.sbs.arizona.edu/~hharley/PDFs/HarleyThesis.pdf

5. Holmberg, A., Nayudu, A., Sheehan, M. (2009). Three Partial Null-Subject Languages: A comparison of Brazilian Portuguese, Finnish and Marathi. Blackwell Publishing Ltd.: Studia Linguistica, Vol. 63 (1), pp. 59-97. DOI: 10.1111/j.1467-9582.2008.01154.x

6. Kinn, K., Rusten, K. A., Walkden, G. (2016). Null Subjects in Early Icelandic, 53 p. Post-print version; published as: Kinn, K., Rusten, K. A., Walkden, G. (2016). Null subjects in early Icelandic. Cambridge: Cambridge University Press, Journal of Germanic Linguistics, Vol, 28 (1), pp. 31-78. DOI: 10.1017/S1470542715000136

7. Langacker, R. W. (2008). Cognitive Grammar: A basic introduction. Oxford: Oxford University Press, x+562 p.

8. Poole, E. (2016). Deconstructing subjecthood. Ms., UMass Amherst, 39 p. http://ling.auf.net/lingbuzz/003197

9. Rosenkvist, H. (2009). Null Referential Subjects in Germanic Languages - an Overview. Working Papers in Scandinavian Syntax, Centre for Languages and Literature, Lund University, Vol. 84, pp. 151-180.

10. Sigurdsson,. H. A. Conditions on Argument Drop. MIT Press: Linguistic Inquiry, Vol. 42 (2), pp. 267-304. DOI: 10.1162/LING_a_00042

11. Taverniers, M. (2005). Subjecthood and the Notion of Instantiation, 46 p. Preprint version; published as: Taverniers, M. (2005). Subjecthood and the notion of instantiation, Elsevier Ltd.: Language Sciences, Vol. 27, pp 651-678. DOI:10.1016/j.langsci.2005.07.003

12. van Gelderen, E. (2013). Null Subjects in Old English. MIT Press: Linguistic Inquiry, Vol. 44 (2), pp. 271-285. DOI:10.1162/LING_a_00127

13. Walkden, G. (2013). Null Subjects in Old English. Cambridge: Cambridge University Press, Language Variation and Change, Vol. 25(2), pp. 155-178. DOI:10.1017/S0954394513000070

14. Walkden, G. (2013). Null Subjects in Old High German and Old Saxon. Paper presented at the workshop Historical Syntax of German, University of Bamberg, 17-18 May 2013, 12 p. http://walkden.space/Walkden_2013_Bamberg.pdf

Размещено на Allbest.ru