THE USE OF THE LEXICAL EXPONENTS OF HYPOTHETICAL MODALITY IN POLISH AND LITHUANIAN

In this article the author focuses on the issue of hypothetical modality in Polish and Lithuanian. A list of the basic exponents of hypothetical modality in both languages is presented. However, the focus is mainly placed on the lexical exponents. On the basis of one of the six groups, which describes a high degree of probability (H5), the differences between the use of the lexical exponents in both languages are examined. In the study, multilingual corpora resources, including The Polish-Lithuanian parallel corpus Clarin-PL., are utilized.


Introduction
Studies on hypothetical modality have been conducted for many years.Recently, however, in the age of the obvious development of parallel corpora, work on the exponents of hypothetical modality specified herein has gained new meaning.The exponents identified earlier can be easily verified, whilst the list of all the linguistic means used for expressing this modality has been considerably extended.It is possible to conduct statistical research and an equiponderant analysis of equivalent expressions in both languages on the basis of multilingual corpora resources.
In this article I refer to my earlier works (i.a.Roszko, D., 2006Roszko, D., , 2011Roszko, D., , 2015) ) and those conducted with Roszko, R. (e.g.Roszko, D. & Roszko, R., 2012), in which I gradually extended and revised the list of hypothetical exponents in the Polish and Lithuanian languages.
The methodological basis for the research conducted on hypothetical modality was taken from theoretical contrastive studies utilizing the interlanguage, which I will discuss in detail below.
The Use of the Lexical Exponents of Hypothetical Modality in Polish and Lithuanian for Polish-Bulgarian Contrastive Grammar, modality is a sentence category.The phrasality of the semantic category results (1) from the orientation of the analytical research conducted from the contents plane to the form plane and ( 2) from an indirect approach to semantics (Maldžieva, 2003, p. 11).The modality itself is inseparable from the notion of a mental state, with which the addresser expresses lack of knowledge of the content conveyed by himself / herself.This mental state, just like the state of the utterance, is connected to the addresser.Both states (utterance and mental) are concurrent.
V. Maldžieva assumes that the mental state always accompanies the utterance state.As a consequence of such an assumption, it is necessary to conclude that the mental state may either correspond to knowledge (indicative modality) or not (non-indicative modality), comp.(Koseska-Toszewa, Maldžieva, & Penčev, 1995;Maldžieva, 2003, p. 12).
V. Koseska-Toszewa does not agree that the mental state is an inseparable element of every utterance.The very introduction of a notion which does not distinguish modal utterances from non-modal ones is a misunderstanding stemming from the perspective of dichotomous division (commonly used in science).Thus, she suggests a modality description within the Petri net discrete theory, based on such elements as fork (splitting), branching and local state (Koseska-Toszewa, Korytkowska, & Roszko, R., 2007, p. 60;Dimitrova & Koseska, 2014, pp. 57-59).
My article assumes that all which is not true (value "1") or false (value "0") is modal, i.e. elements to which the value of truth or the value of falsity cannot be assigned.In order to demonstrate this, I will quote the following example: The sentence Jan zdał wczoraj egzamin z języka angielskiego.'Yesterday John passed his English exam' may have the value of truth or falsity, because this sentence has two mutually exclusive interpretations.Either the addresser's utterance is true or false.That is why I consider such sentences to be non-modal.The following sentence is an example of a modal sentence: Chyba Jan zdał wczoraj egzamin z języka angielskiego.'Yesterday John may have passed his English exam.'This sentence cannot be assigned the value of truth or the value of falsity.That is why I conclude that this mental state is correct only for modal sentences, comp.(Koseska-Toszewa et al., 2007, p. 500), where two types of modality are distinguished: possibility (containing the functor of possibility) and obligatory (containing the functor of obligation).In this article we will focus on the modality of possibility, especially on the semantic category of hypothetical modality.

The semantic category of hypothetical modality
As I have mentioned above, we distinguish two kinds of modality: one with a possibility functor (it is possible that) and one with an obligatory functor (it is necessary that).Hypothetical modality is one of the modal categories containing such a possibility functor.We can also give examples of other semantic categories containing such possibility functors, such as conclusivity or imperceptibility.According to the theoretical assumptions of contrastive studies, the semantic category of hypothetical modality is a sentence category.The contents of the semantic category of hypothetical modality present the addresser's subjective attitude to opinions uttered by himself/herself, comp.(Maldžieva, 2003).Maldžieva distinguishes six degrees used to express probability.She assumes that the first degree (H11 ) has the lowest probability and degree six (H6) has the highest.The degrees from second to fifth are positioned between the extreme degrees (first and sixth), see Figure 1.
As can be observed in Diagram 1, the probability value rises concurrently with the increase in the degree numbering of the hypothetical H units.According to the assumptions of the theoretical contrastive studies utilizing the semantic interlanguage, used in the description of modal categories, The Use of the Lexical Exponents of Hypothetical Modality in Polish and Lithuanian ["0"] -<H1 -H2 -H3 -H4 -H5 -H6> -["1"] "0" -false "1" -true Figure 1: Hypothetical modality.Position of the respective probability degrees on the axis V. Maldžieva rejects the extreme probability values, i.e. falsity ("0") and truth ("1").I agree with this claim, which is shown in Diagram 1 by isolating the range of the individual probability values <H1 . . .H6> from the extreme values of falsity and truth.Moreover, V. Maldžieva eliminates intermediate values between falsity and truth, for which the arbitrary value "½" can be accepted.As it can be seen in Diagram 1, I have not rejected this value.Accepting the "½" value differentiates my approach from that of V. Maldžieva's.The analysis of languages such as Lithuanian shows that the morphological exponents do have the "½" value, which I describe in a monograph discussing the functions of perfectum in the Lithuanian language and the Lithuanian dialect of Puńsk (Roszko, D., 2006).

Hypothetical modality exponents in Polish and Lithuanian
On the basis of the research conducted in (Roszko, D., 2015) I conclude that in both Polish and Lithuanian the typical exponents of hypothetical modality are lexemes (e.g.Pol.może 'maybe', Lit.gal 'maybe').Some of the lexemes may cause the syntax to become subordinate/dependant in nature (e.g.Pol.Wydaje mi się, że . . .'Seems to me that ...', Lit.Man atrodo, kad . . .'Seems to me that . . .'), whereas other lexemes may function as interpolation (e.g.Pol.widać 'as it appears', lit.matyt 'as it appears').Furthermore, in the Lithuanian language we can observe morphological exponents, including ones based on modus relativus, described in the Grammar of the Lithuanian language (Ambrazas, 1997), as well as on forms of the Future Perfect Tense, subjunctive structures etc., whereas in the Polish language we can observe paramorphological structures (type of musiał (musi ) + bezokolicznik 'może', powinien + bezokolicznik 'może', mógł + bezokolicznik 'może') On the basis of the bilingual Polish-Lithuanian corpora (Experimental and Clarin-PL) it can be concluded that part of the Lithuanian lexical exponents may differ from the forms known in Polish, used unconventionally with this meaning, e.g.: • forms of the imperative mood of the verb sutikti 'to agree': sutik (sg.) [literally: agree], sutikite (pl.) [literally: agree], • or 2 nd person future form of the verb rasti 'find; come upon; find out; learn; meet; come across"rasi [literally: you will find] used as interjection.
It is worth noting that in the Lithuanian language the combination of the negative and positive forms is used more often than in Polish (e.g.brolis ne broils 'maybe a brother' [literally: a brother not a brother]) as a hypothetical exponent of group H4.A characteristic feature of the Polish language is the use of the improper intransitive verb (predicative) widać.
I would like to examine the lexical exponents in greater detail.That is why, without going into detail as far as non-lexical exponent structures are concerned, I will proceed to the description of lexemes, common in both languages, in the hypothetical modality function.In Polish and Lithuanian, the lexical exponents create a clear structure whose main feature is, in addition to expressing the very content of hypothetical modality, an indication of the degree of probability.The concept of distinguishing six groups (from H1 to H6) has already been presented above, and in which the group of lexemes described by H1 expresses the lowest degree of probability (close to the logical value of "0" / falsity), and each successive one respectively expresses a higher level of probability.The group marked H6 is assigned the highest level of probability (close to the logical value "1" / truth).

The Use of the Lexical Exponents of Hypothetical Modality in Polish and Lithuanian
In Table 1 2 I present examples of the lexical exponents which express the contents of hypothetical modality with division into groups.For more on this see (Roszko, D., 2015).The division into groups is not precise.It needs to be emphasized that verifying this very feature is possible due to analysis of the multilingual corporal resources.Individual lexical exponents may represent a small proportion of their use attributed to expressing the degree of probability that is proper for the neighbouring group.For instance, the exponent of group H4 may be occasionally used instead of the exponent of group H5.Then the same exponent of group H4 (appearing in the place reserved for the exponent of group H5) cannot replace the exponent of group H3.Therefore, co-sharing part of the exponents for the neighbouring groups is only possible unilaterally, i.e. exponent a assigned to group H4 can be used to express the degree of probability typical for groups H3-H4, whereas exponent b of the same group H4 can be used to express the probability degree typical for groups H4-H5 etc.
However, on the basis once more of the two Polish-Lithuanian corpora, it has to be mentioned that some exponents dissent from the general rule.These are: (1) the Lithuanian matyt 'może' whose use has been noted for three groups-H4-H5-H6 and (2) the Lithuanian gal, which seems to have an even broader scope.It needs to be emphasized that we get such a result only through a concurrent analysis of two paralleled language resources.If the analysis of these Lithuanian lexemes was conducted using a monolingual resource, then the Lith.form gal 'może' might be considered a neutral exponent of hypothetical modality even-handedly with the morphological exponent, i.e. modus relativus forms.
In Table 2 I present statistic data which describe individual groups of exponents of hypothetical modality on the basis of the Polish-Lithuanian parallel corpus Clarin-PL (Roszko, D. & Roszko, R., 2016) and the Experimental Polish-Lithuanian corpus.I have divided the data into two groups.The first group contains a number of unique lexemes in individual groups.The second includes the total number of noted appearances of these forms, also divided into groups.The total volume of both of the corpora, used for defining the particular degrees of probability, amounts to 8 million words.
Analysis of the data contained in Table 2 (columns 2 and 3) confirms low exponent representation in groups H1-H3.Then, an increase in the exponent number in group H4 can be observed.As can be clearly seen, the climax is reached in group H5, and then the number of unique exponents drops (noticeably for the Polish language) in group H6, although it remains high.Analysis of the data contained in Table 2 (columns 3 and 4) shows that in groups H1-H5 the frequency of the exponent use rises in tandem with the rise in probability degree.However, for group H6 we can observe a decrease in frequency compared to group H5 (for both languages).Nevertheless, the lexemes of groups H5 and H6 are the most frequently used lexical exponents by both groups of languages users.A lower total frequency of the Lithuanian exponents of hypothetical modality certainly draws attention.However, it is not a result of hypothetical contents being less frequently exposed in the Lithuanian language (compared to Polish).Instead, it is a result of the common use of morphological exponents in the Lithuanian language.
Similar conclusions can be drawn on the basis of the total number of unique lexemes in both languages.Namely, the Polish language with a total of 88 lexemes seems to be more diverse than the Lithuanian language with a total of 72 unique lexemes.
Not by chance do I mention in the paragraph above that the Polish language seems to be more diverse than the Lithuanian language, because a detailed analysis of the corpora resources presents new facts, which make us look at both languages from a different perspective.
5 Analysis of the hypothetical modality exponents, included in group H5, in Polish and Lithuanian The choice of group H5 for a detailed parallel Polish-Lithuanian analysis of lexical exponents in both languages is not accidental.Firstly, in group H5 we can observe a large number of lexical units: 49 Polish lexemes in group H5 and 39 lexemes in the remaining five groups (H1, H2, H3, H4 and H6).Secondly, there is a high frequency of lexemes in group H5 in Polish: 4613 lexeme uses in this group against a total of 2869 of all the lexemes included in the remaining groups.
There is an analogous situation with the Lithuanian language: 3017 against 2381.Thirdly, there is a significant difference in the frequency of use between the Polish and Lithuanian lexemes of the very same group (H5): 4613 (Polish exponents) against 3017 (Lithuanian lexemes).Numerical data is contained in Table 3.
The group H5 exponents express a high degree of probability, much higher than value.On the basis of excerption of the two aforementioned corpora, the following exponents can be put into group H5 (after each exponent we see the number of uses in the two corpora):  LT Ir žinote, man atrodo, kad šičia visai neblogai.
The examples above have been selected in order to present the larger diversity of exponents in the Lithuanian language.In examples 1-3, the Polish lexeme może has a corresponding lexeme in the Lithuanian language: in the first sentence it is galbūt, in the secondgal, and in the third it is galimas daiktas.However, in the Lithuanian sentence 4 there is no lexeme directly corresponding to the Polish może.Nonetheless, we can note a form of interrogative sentence, which in an equivalent way reflects the addresser's uncertainty in relation to the contents conveyed.In sentence 5 we see the Polish and Lithuanian lexical exponents which cause the syntax to become subordinate in nature: Pol.uważam, że (I think that) . . .and Lith.man atrodo, kad . . .

Statistical analysis of the hypothetical modality lexical exponents in Polish and Lithuanian, included in group H5
Firstly, it is worth ranking the above-mentioned exponents in descending order of their frequency in the corpora.It is certainly surprising to find the Polish exponent może in the top position, with a disproportionally high number of appearances against other exponents, amounting to 2261.As a reminder, in this group 49 unique exponents have been noted, with a total of appearances amounting to 4613.While analysing the numerical data contained in Table 4 it is necessary to remember that the total number of uses of any hypothetical modality exponent in group H5 in the Polish language amounts to 4613, whilst in the Lithuanian language it is substantially lower and amounts to only 3017 (it is noted that for 65% of the Polish exponents appearances an equivalent Lithuanian lexeme is used).Despite this, it can be observed that with the first seven exponents, 5 of them are Lithuanian forms (gal, žinoma, atrodo, tikrai, galimas daiktas) and only 2 are Polish (może, The Use of the Lexical Exponents of Hypothetical Modality in Polish and Lithuanian oczywiście).In the Polish language the most dominant exponent is może, which appears 3.3 times more often than the second most dominant Polish exponent, and 14 times more often than the third most dominantbyć może.As regards the Lithuanian language, these proportions are significantly lower, i.e. the most dominant exponent gal is noted 2.1 times more often than the second most dominant žinoma and 4 times more often than the third most dominant Lithuanian exponent atrodo.
The comparison of the total number of appearances of Polish and Lithuanian exponents, although very informative, does not fully reflect the facts.This is because the disproportion of the total number of uses of these lexemes in both languages is quite distinct.(comp.4613 appearances of Polish lexemes to 3017 appearances of Lithuanian lexemes).That is why it has been concluded that presenting the percentage participation will better illustrate this Polish-Lithuanian disproportion in using the individual lexemes.The percentage values have been calculated on the basis of uses of a given lexeme against the total number of uses of all the lexemes (in group H5) in a given language.The results are presented in Table 5 5 .
The analysis of the usage of each exponent in comparison to the usage of all the group H5 exponents in a given language changes the order of the most frequent lexemes.The Polish exponent może still remains in the highest position with its 49.01%participation in the Polish H5 group.The second highest position goes to the Lithuanian gal with its 35.50% participation in the Lithuanian H5 group.The third most frequent exponent is the Lithuanian žinoma with its participation of 16.74%, which replaces the Polish oczywiście, noted in this position in Table 4.The Polish oczywiście occupies position 4 with its participation of 14.72%.Positions 5, 6 and 7, like in Table 4, are taken by the Lithuanian exponents atrodo, tikrai, galimas daiktas, which respectively have participations of 8.85%, 8.35% and 5.60%.A significant fact is the drop of the Polish być The Use of the Lexical Exponents of Hypothetical Modality in Polish and Lithuanian może to position 10 (the participation of this exponent is 3.49%).It has been overtaken by two successive Lithuanian exponents, savaime aišku and galbūt, with respective participations of 5.60% and 4.24%.In summary, within the top ten most frequently used exponents we can find 3 Polish forms and 7 Lithuanian forms.
On the basis of the total number of unique exponents noted in both languages, we can define the average participation of the exponent in its group.As regards the Polish language, the average participation is 2.04%, and for Lithuanian it is 3.23%.When comparing these values against the participation of the individual exponents, we can conclude that in the Polish language the considerable domination of one exponent can be observed: (może -49.1%).This allows us to draw the conclusion that practically every other exponent of hypothetical modality used for expressing the fifth degree of probability (H5) is the Polish może.When we take into account the respective participations of the remaining Polish lexemes and at the same time look at the facts registered in the Lithuanian context,6 it is necessary to conclude that the diversity of Polish exponents is low.Of course, the total number of Polish lexemes is considerably higher than their Lithuanian counterparts, but their distribution is apparently uneven.

Conclusions
Using Polish and Lithuanian as an example, it can be concluded that hypothetical modality may be expressed by lexical (including simple and complex expressions) and morphological means.In the Polish language the dominant exponents are lexemes.In the Lithuanian language there is a balance between the use of lexical and morphological exponents.This is why, in a numerical approach to the use of lexical exponents in both languages, a more frequent hypothetical marking of a sentence can be observed as a result of using the lexeme in the Polish context (comp.group H5 described herein: 4617 uses of Polish lexemes corresponding to 3017 uses of the Lithuanian lexical counterparts).Moreover, a larger number of unique lexemes in the Polish language is observed (for group H5 it amounts to 49 unique Polish lexemes and 31 Lithuanian counterparts).
However, a detailed analysis of the corpora resources reveals a disproportion between the total number of the lexemes and their frequency.It appears that in the Polish language only one lexeme (Pol.może) is noted in every second hypothetical sentence of group H5.Almost all the other lexemes are used much less frequently.In the Lithuanian language we do not observe an explicit domination of one of the exponents.Most of the Lithuanian hypothetical exponents have a higher frequency, which is followed -compared to the Polish language -by a lexeme diversity in the hypothetical function.

Table 1 :
Selected lexical exponents of hypothetical modality in Polish and Lithuanian The Use of the Lexical Exponents of Hypothetical Modality in Polish and Lithuanian

Table 3 :
Lexical exponents of hypotheticality in Polish and Lithuanian in a statistical approach.Group H5 vs remaining groups (on the basis of the Polish-Lithuanian parallel corpus Clarin-PL(Roszko, D. & Roszko, R., 2016)and Experimental Polish-Lithuanian Corpus)The Use of the Lexical Exponents of Hypothetical Modality in Polish and Lithuanian 5. PL I wie pan co, uważam, że tu jest zupełnie nieźle.

Table 4 :
Hypothetical modality lexical exponents in Polish and Lithuanian.(group H5) with a frequency above 10 appearances in the Polish-Lithuanian parallel corpus Clarin-PL (Roszko, D. The Use of the Lexical Exponents of Hypothetical Modality in Polish and Lithuanian

Table 5 :
Percentage participation of the hypothetical modality lexical exponents, presented in table 4, in Polish and Lithuanian (group H5) The Use of the Lexical Exponents of Hypothetical Modality in Polish and Lithuanian