FINITENESS AND NON-FINITENESS IN FINNO-UGRIC LANGUAGES. SYNCHRONIC, DIACHRONIC AND CROSS-MODULAR ISSUES

Abstract

Kristiina Ross suggested that since the identical morphemes used in the Estonian and Finnish infinitival systems occur in different combinations and syntactic patterns in the two languages, it can be hypothesized that the system of Estonian infinitival constructions rose relatively late and is likely to be influenced by foreign sources and by the linguistic ideas of the 16th- 18th century Bible translators. The authors showed however that the meanings of the constructions do not derive from the presence or absence of the Agent but from the inherent meanings of the participle: the impersonal meaning deriving from the impersonal paradigm of which the participle is part, and the aspectual (resultative ~ perfective) meaning, which seems to be an additional (new) meaning and may be related to the Germanic resultative participle through language contact. The distribution of the genitive agent is subject to various restrictions: it combines only with non-finite verb forms, it must immediately precede the verb form and it does not co-occur freely with other VP constituents. The combination of the genitive agent and the verb form thus poses a descriptive, theoretical and explanatory challenge: its distribution is neither fully general nor lexically constrained, it is not describable as a regular VP, the syntactic role of the genitive NP is difficult to define, and the information structure of the phrase is unusual.

Finno-Ugric languages, especially the Finnic branch, have a rich non-finite verb system. The grammar and semantics of non-finite forms and finiteness in general raise a series of synchronic and diachronic issues. In order to discuss them, a seminar was organized by the Institute of the Estonian Language and the foundation Fenno-Ugria in 2009. The idea to organize the seminar stemmed from discussions over the theoretical interpretation of a construction in Old Written Estonian, which was the subject of Kristiina Ross's presentation "The problems of non-finiteness/nominalization in a 17th century construction". In the manuscript of Johannes Gutslaff's Bible translation dating from the middle of the 17th century, the construction [mainfinitive + postposition tarbis 'for'] was used in the same (purposive) function as the present-day translative of the mainfinitive. The presentation raised three questions: 1. Should the -ma form in this construction be interpreted as an infinitive or as a substantive, and what is the use of such a distinction in the first place? 2. Is it possible that the use of this construction was influenced by one of the source languages (Hebrew, Greek, Latin or German)? 3. Was Johannes Gutslaff's usage of the construction supported by the popular language or was it a theoretical construct?

On the one hand, a rich inventory of non-finite forms is a characteristic property of Balto-Finnic languages. On the other hand, Marja Itkonen-Kaila (1997) remarks that the large number of nominal verb forms make Finnish, unlike Swedish and German, similar to the classical languages. Kristiina Ross suggested that since the identical morphemes used in the Estonian and Finnish infinitival systems occur in different combinations and syntactic patterns in the two languages, it can be hypothesized that the system of Estonian infinitival constructions rose relatively late and is likely to be influenced by foreign sources and by the linguistic ideas of the 16th- 18th century Bible translators.

Another talk addressing the issue of the distinction between non-finite verbs and deverbal nouns was Jaakko Leino's presentation "Infinitives and/or deverbal nouns in Finnish". He started out by proposing a view of the Finnish infinitive system according to which there is a single infinitive with a paradigm of case forms. The forms are in complementary distribution and occur in a limited set of specific constructions. As a generalization across these forms, he proposed an overall infinitive construction in construction-grammatical and cognitive-grammatical terms, and went on to contrast this with deverbal nouns. Jaakko Leino argued that the received wisdom about infinitives as noun-like verb forms that are modified like verbs, and deverbal nouns as verb-like nouns that are modified like nouns, faces the following problems. 1. The distinction between the deverbal noun in -minen and the identical-looking 4th infinitive; this distinction looks counterintuitive to many native speakers. 2. The genitiveaccusative syncretism which blurs the distinction between the object, the genitive subject and the genitive attribute. 3. The lack of knowledge of the non-standard language and the variation to be found there, i.e. of the extent to which the infinitive system is an artefact of the standard language. 4. The lack of knowledge of different deverbal noun types.

The deverbal nouns may form a similar coherent system of division of labor as the infinitives do. In order to tackle these problems, Jaakko Leino posed the need to study non-standard language and dialect syntax, to examine the deverbal noun types both individually and possibly as an overall system, and to conduct a close comparison of the modification properties of (different types of) deverbal nouns and infinitives.

Pille Penjam's talk "The constructions of da- and ma-infinitives in Old Written Estonian" compared the uses and functions of infinitival complements in Old Written Estonian and contemporary written Estonian, using corpus data from the 17th century (a selection texts by Georg Müller, Heinrich Stahl and Christoph Blume, approx. 290,000 words) and the 20th century (morphologically disambiguated subcorpus of the Tartu University Corpus of Written Estonian, approx. 215,000 words). As a characteristic of the 17th century texts, Pille Penjam pointed out the extensive free variation of da- and ma-infinitives. Overall, the relative frequency of da-infinitive constructions was considerably lower in 17th century than in 20th century (883.1 versus 2277.2 instances, per 100,000 words), whereas the relative frequency of ma-infinitive constructions was higher (1283.4 versus 854.9 instances). The contemporary Estonian da-infinitive constructions tend to lack counterparts in Old Written Estonian. The most frequent were the da-infinitive constructions functioning as objects and expressing a wish or intention, and the modal construction with võima 'can'. The dominant mainfinitive constructions were the modal construction with pidama 'must' and the future construction with saama 'get'. The 17th century data contained some uses of infinitives with no ma- or da-infinitive counterparts in contemporary Estonian. Georg Müller's language appeared closest to the infinitival government in contemporary Estonian.

The subject of the seminar was addressed in the most general way by Östen Dahl in his talk "Finiteness and non-finiteness". He argued that a universal definition of finiteness in terms of specific morphological categories (person, number, tense, mood) is impossible since the categories vary from language to language, and even within one and the same language properties commonly associated with finiteness do not always occur together. He proposed not to focus on finiteness as a property, or as a distinction between finite and non-finite, but to search for finiteness-related phenomena (FRP), i.e. grammatical markings, rules and principles that are typically obligatory in the prototypical contexts, active affirmative declarative clauses, but which may also be extended to other (but not all) constructions.

Core finiteness-related phenomena are inflectional markings, which are grammaticalized in most languages of the world: person marking (above all subject marking), tense, aspect (mainly perfectivity/imperfectivity), evidentiality, negation and politeness; it is also tempting to see obligatory subject pronouns as a kind of periphrastic finiteness marking. It is difficult to define precisely what is the function of these markings, but it might be significant to look where they occur: either on main verbs or on auxiliaries but also in Wackernagel's position, i.e. the position after the first major constituent of the sentence, which, in many languages, hosts auxiliaries and clitic pronouns (and, in Germanic, also finite verbs), particularly in main clauses.

Different FRP may vary in where they occur, but they always show up in prototypical FRP contexts. Contexts where FRP tend to be reduced are sentences with nominal predicates, imperatives, negated sentences and questions. The fact that the prototypical combinations can be seen to be active verbs + imperative and nominal predicates (non-active) + declarative, raises the interesting possibility that FRP prototypically mark the non-prototypical combination active verb + declarative. Examples of diachronic processes that are frequently the origin of the asymmetries between finite and non-finite clauses: the inflectionalization of derivational morphemes/processes, the grammaticalization of markers starting out in main clauses and the recycling of non-finite forms as finite forms through insubordination.

The development of non-finite forms into finite forms was addressed in more detail by Iris Metsmägi's presentation "Verb forms of non-finite origin in personal paradigms of Finnic, Saami and Volgaic languages". The forms in question are originally deverbal nominals and have still preserved their nominal character in that no personal endings are added to them. Two different types of personal paradigms involving forms of non-finite origin can be distinguished: 1. Both finite forms (forms with personal endings) and forms of non-finite origin occur in the paradigm. A form of non-finite origin mainly occurs as the 3P of the present indicative (in Finnic languages in most cases; in Mordvin languages; in Mari more restrictedly), but it can be a 3P form of some other tense as well (e.g. the Pl3P of the Hill Mari past I); there also exist inflectional paradigms that include more than one form of non-finite origin (e.g. the indicative present of Livonian and Saami). 2. A form of nonfinite origin forms the whole paradigm (the same form without personal endings is used for each person, plural marker may be added), as in the Quotatives of Estonian and Livonian. This is a different type of paradigm than the old ones: person is not marked by personal endings but by personal pronouns. New finite categories may thus lead to a new type of paradigm.

Tatíjana Agranat's presentation "On the voice oppositions in the Balto-Finnic Languages" studied Estonian, Finnish, Karelian, Veps, and Votic non-finite complements. The examples contained inessive non-finite forms, t- and m-formative non-finite forms (supines). In her analysis, the minimal pairs of non-finite forms display voice oppositions (as in the Estonian sentences Mees on hea kündma 'The man is good at ploughing' vs. Põld on hea künda 'The field is easy to plough'), and the choice of a non-finite form depends on the semantic role of the argument which correlates with this form.

Three talks addressed the semantics of particular non-finite forms in Estonian. Ilona Tragel and Liina Lindström's presentation "Estonian past passive participle" discussed the semantics of the past passive participle. The constructions that contain it have two basic meanings, impersonal and resultative. The meaning of the construction correlates with the occurrence of an explicit Agent in the construction: constructions with a low possibility to use the explicit Agent are impersonal (ole 'be' impersonal and saa 'get' impersonal), whereas constructions with a high possibility to use the Agent are aspectual (perfective) (ole- and saapossessive perfect and saa-resultative), with the passive construction in the middle. The authors showed however that the meanings of the constructions do not derive from the presence or absence of the Agent but from the inherent meanings of the participle: the impersonal meaning deriving from the impersonal paradigm of which the participle is part, and the aspectual (resultative ~ perfective) meaning, which seems to be an additional (new) meaning and may be related to the Germanic resultative participle through language contact. The authors also discussed the development of the possessive perfect.

Renate Pajusalu and Heili Orav's presentation "Asymmetry in expressing motion event: Estonian supine construction" examined how supine constructions are used to encode the spatial characteristics of motion events, starting from the hypothesis that there is an asymmetry in favour of the encoding of the goal. They compared the usage frequencies of the illative form of the supine, which encodes the goal, the inessive form of the supine, which encodes the location, and the elative form of the supine or the mast-supine, which encodes the source. The study confirmed the frequency-bias toward the goal, but it was surprising that in the case of the supines the bias was much stronger than in the locative NPs. In their presentation, the authors also discussed the difficulties related to the semantic interpretation of the supine forms in context and the problems of teasing apart the spatial, temporal and activity meanings of the supines.

Anne Tamm argued in her presentation "The absentive in Estonian" that the study of implicatures and presuppositions may help to distinguish the meanings of the absentives cross-linguistically. Caspar de Groot (1995) proposed the term absentive for constructions that signal the absence of its subject from the deictic center. The inessive m-formative non-finite form (e.g. Jaan on ujumas 'Jaan is off swimming') is tacitly assumed to be an instance of the absentive (Vogel 2007; Tommola 2000). Similarly to several European languages, the Estonian absentive has the following characteristics: 1) the referent of the subject is absent, 2) the Subject is involved in an activity, 3) it is predictable how long the Subject will be absent, 4) the Subject will return after a period of time. These meanings are implicatures in Estonian and the Estonian absentive does not meet several standard conditions. Most importantly, the subject and the speaker do not have to differ in their location in Estonian in order for the absentive meanings to emerge, cf.: Kus Jaan on? Ta pidi ju kell neli siin mind ootama 'Where is Jaan? He was supposed to wait for me here at four'. Ta on siin minu kõrval ujumas 'He is off swimming here next to me.' The Estonian inessive-based absentive has developed from a typical communicational situation where locative expressions are used: answering the question about the whereabouts of somebody who is absent from the deictic center. This situation has yielded the obligatory presupposition of expected presence of the subject in the deictic centre in the nonfinite - but not simple NP - locatives in case of the Estonian absentive.

Heete Sahkai's presentation "The genitive agent in Estonian non-finite constructions" examined an idiosyncratic piece of Estonian non-finite syntax, the genitive NP realizing the actor argument of nonfinite verb forms, as in Otsus on Peetri tehtud 'The decision has been made by Peter'. The distribution of the genitive agent is subject to various restrictions: it combines only with non-finite verb forms, it must immediately precede the verb form and it does not co-occur freely with other VP constituents. As a further exceptional property, the genitive agent tends to be focal when the verb form functions as the semantic predicate of the clause; this is in contrast to the more usual postverbal focus position of the Estonian VP. The combination of the genitive agent and the verb form thus poses a descriptive, theoretical and explanatory challenge: its distribution is neither fully general nor lexically constrained, it is not describable as a regular VP, the syntactic role of the genitive NP is difficult to define, and the information structure of the phrase is unusual. As a possible explanation to this idiosyncracy, Sahkai proposed the status of the genitive agent as a historical relic and posed the need to study its diachronic origins.

References

REFERENCES

Groot, C. de 1995, The Absentive in Hungarian. - Levels and Structures, Szeged (Approaches to Hungarian 5), 45-61.

Itkonen-Kaila, M. 1997, Mikael Agricolan Uusi testamentti ja sen erikieliset lähtötekstit, Helsinki.

Tommola, H. 2000, Progressive Aspect in Baltic Finnic. - Tense and Aspect in the Languages of Europe. Empirical Approaches to Language Typology. EUROTYP 20-6, Berlin-New York, 655-692.

Vogel, P.M. 2007, Anna ist essen! Neue Überlegungen zum Absentiv. - Kopulaverben und Kopulasätze. Intersprachliche und intrasprachliche Aspekte, Tübingen (Linguistische Arbeiten 512), 253-284.