EXPLAINING SYNONYMY BETWEEN OBLIGATION EXPRESSIONS IN FINNISH LEGAL LANGUAGE – RESULTS OF A SURVEY

The aim of this research is to examine synonymy between two obligation expressions, tulee and on -t(t)ava, in Finnish legal language on the basis of the results of a survey. The analysis is based on the responses of 336 Finnish language users. Both expressions are the most frequent means of expressing obligation in normative acts and they can be regarded as nearsynonyms. However, the synonymy between them has not been studied in legal language so far. On the basis of respondents’ comments ten differentiating features have been found. The features distinguished in the research can provide some additional information on the usage of the expressions because not all of the features have been addressed in the dictionaries so far.


Introduction
Investigating deontic modality in Finnish and Polish legal language firstly requires deep understanding of the semantic relations between deontic expressions in each language separately.It is significant to choose the right equivalent of a deontic expression due to a possibly different established means of indicating modality.This refers especially to the domain of translations (Vanden Bulcke 2013: 15).As meanings of obligation expressions are very close to each other and the existing dictionary entries appear insufficient to discriminate between them, some other differentiation seems crucial.
The aim of this article is thus to identify some differences in meanings of two near-synonymous obligation expressions in Finnish legislative texts, on -t(t)ava and tulee by assigning them differentiating features.Ten features have been obtained from a survey that was conducted among Finnish natives and will be discussed further in the article.It will include an analysis summarising the results obtained from all respondents and additionally basing on the educational profile of three groups of respondents.

Notions on synonymy
Synonymy is based on an idea that one concept can be denoted by a couple of linguistic means.Substitution is crucial for explaining this phenomenon.This means that some words or phrases can be regarded synonymous when their substitution does not change the truth value of the proposition (Strazny 2005).If they cannot be intersubstituted in all contexts without losing or changing a part of their meaning, they are not absolute synonyms, but rather near-synonyms, quasi-synonyms or 'plesionyms' in Cruse's terms adapted by Edmonds and Hirst (2002:107;116).Near-synonyms can vary in "their shades of denotation, connotation, implicature, emphasis, or register" (Edmonds and Hirst 2002:107 after DiMarco, Hirst, and Stede 1993and Goźdź-Roszkowski 2013: 96).Moreover, in regards to a legal text, it is the contextual dimensiona specific domain and genrethat plays a significant role in explaining near-synonyms (Goźdź-Roszkowski 2013: 108).
In legal context the analyses of synonymy often concern legal phraseology (Chromà 2011) rather than deontic expressions.Deontic expressions have been rather studied in terms of textual fit (Biel 2014), parameter-based approach being a theoretical description of deontic units (Nowak 2011) and functional equivalents for translation purposes (Matulewska and Gortych 2009).
In general and standard Finnish language deontic modality is a more deeply studied subject, including analysis of informal, spoken language, where these means are used very frequently (Flint 1980, Kangasniemi 1992).However, the usage of Finnish deontic expressions in the legal language has been scarcely described, neither from the point of view of synonymy (Kangasniemi 1986, Kanner 2011, Attila 2016), nor has it been raised much in Finnish legal theory.1One of the motives of the author was therefore to analyse the subject further.This analysis was also inspired by some Finnish studies on verbal synonymy, although they did not deal with legal genre.They are Ulla Vanhatalo's research on variations of the verb valittaa 'to complain' in informal language (Vanhatalo 2002) and Antti Arppe's findings on differences between miettiä and pohtia, 'to think' and 'to consider', to name just a few (Arppe 2007).

Two deontic expressions in the analysis
The subject of this analysis are two expressions, on -t(t)ava and tulee.According to a Finnish-English dictionary, they mean 'must', 'have to', 'have got to'.Tulee is a form of infinitive tulla in the 3 rd person singular.The non-modal meaning of tulla is 'to come'.The necessity expression on -t(t)ava consists of the infinitive olla 'to be' in the 3 rd person singular and a passive participle of the complement verb.As both expressions convey necessive meaning, they get a genitive subject.Other modal verbs that are not necessive have a nominative (e.g.modal verbs of possibility).It is a characteristic feature for obligation expressions in Finnish (Laitinen 1993: 11).
The choice of these two expressions is motivated by their frequency.On -t(t)ava is the most frequent exponent of necessity in written general Finnish, with tulee on the third place.The second and the fourth place is taken by aforementioned modal verbs, pitää and täytyy (Kangasniemi 1986: 84).What's interesting, on -t(t)ava is also the most frequent obligation expression in the language of Finnish legislation, and tulee is the second most common obligation expression, although with much lower frequency (Kanner 2011: 34-35).This was confirmed in author's study.In author's corpus of fifty most often used legislative texts (as on September 7, 2015) gathered from an online database Finlex the analysed expressions on -t(t)ava and tulee were most frequent amongst other obligation expressions and the colloquial ones pitää and täytyy were insignificant and performed normative functions only in single cases (Table 1).The expressions can be considered as near-synonyms.There is no absolute synonymy between them, although entries in a Finnish-

Obligation expression
Frequency in the corpus on -(t)tava 6668 tulee 803 täytyy 4 pitää 1 English dictionary do not differentiate them from each other in the deontic context at all, in both cases referring to the other necessive verb täytyä, which is used in the everyday language and means 'must', 'have to', 'have got to'.However, there is a significant difference in their frequency in legislative texts as shown above, and their definitions regarding the deontic meaning in a Finnish-Finnish dictionary are not identical but, instead, adduce some differentiating nuances of meaning (Table 2 and 3 The same definitions in English: Table 3. Dictionary definitions of on -t(t)ava and tulee (Source: Kielitoimiston sanakirja 2016, author's translation).
These definitions come from a dictionary of general Finnish language.What is common is that in both definitions it is referred to 'necessity' (välttämättömyys) and 'compulsion' (pakollisuus) (Table 3).The additional element in the definition of tulee is 'to be appropriate, suitable, necessary to do something' which may refer to some kind of recommendation.It is an additional connotation.In both dictionary definitions there is also a reference to modal verbs täytyy and pitää which both mean 'have to' and are also in the 3 rd person singular.They belong to the informal language register, although they can also appear in more formal register, like president speeches.The definitions are therefore insufficient in helping discriminate between obligation expressions for the purpose of using them in legal texts.In order to name some more differentiating nuances between the expressions a survey was conducted among Finnish native speakers to study empirically their perception of the usage of these expressions in legal language.

Methods
An online questionnaire survey was designed by the author at the Institute for the Languages of Finland (Kotimaisten kielten keskus, Kotus) as a part of CIMO scholarship and performed for the purposes of author's doctoral thesis.The survey was available on the Internet for over a month (10 November -16 December 2015).
The concept of a questionnaire survey (kyselytesti) was adopted from Vanhatalo's research (2003) who investigated semantic nuances between a pair of adjectives (keskeinen and tärkeä 'central' and 'important').She compared two methods of analysing near-synonymy: corpus linguistics and population tests, by using a previously done corpus-linguistic study of Janttunen ( 2001) and reviewing its results by involving survey participants to outline the differences.
The survey on the expressions in legal language analysed in this study contained questions about obligation and possibility verbs but in this article the focus is only on the obligation expressions.There were ten sentences presented with gaps in which the deontic expressions were to be chosen from a list of six different options.There was also a possibility to write one's own expression.For each of the sentences respondents could comment on the justification for their choice of particular expressions.The task used thus an idea of substitution.
Ten sentences were excerpted directly from a corpus of 50 most often used (as on September 7, 2015) authentic legislative texts in Finlex, which is a database of Finnish legislation online.The sentences were varied according to different subject-specific domains.First, the corpus was analysed in order to rank the most frequent collocations of the obligation expressions.Then, example sentences were gathered on the basis of the most frequent subjects and verb predicates with the expressions in question.
This analysis is a qualitative study of the respondents' comments.Respondents justified their choice of a certain expression in each question by naming various features and associations.Each feature outlined in these comments was subsequently assigned to one of 10 broader categories, whose more detailed description may be found in the following sections of the article.Created categories and subcategories of features served to create profiles of characteristics.Eventually the profiles were compared to each other on the basis of the distinguished features.It is worth noting that the analysis does not aim at giving exhaustive explications.It rather shows, how Finnish natives, so the language users, perceive the differences between obligation expressions.Therefore all these categories are only based on Finnish respondents' opinions which were examined and classified.
Altogether 336 respondents participated in the survey.The target group concerned mainly persons with legal and language-related background, i.e. whose either education or work, or both, were connected to these areas.However, people with other background were also welcome to participate.From all 336 respondents, the ratio of women to men amounted to approximately 3:1 (women: 74.7%, men: 25.3%).65.5% of respondents were 40 years old or younger.The majority (69.9%) completed higher education.According to the declared completed or ongoing educational profile, areas of law and language were represented by 46.1% and 40.8%, respectively.Also, 40.8% and 41.4% of respondents declared their professions as related to, respectively, the areas of law and language.
While the survey was completed by 336 respondents, some of them had to be excluded from the study.The exclusion criteria were providing no comment for any of the sentences ('empty') or not mentioning any certain feature but only respondent's preference ('preference').Hence, respondents included in the analysis were those who mentioned at least one feature of an expression (Table 4).Finally, for tulee and on -t(t)ava there were, respectively, 146 and 159 respondents included in the study.Being aware of a commonly acknowledged fact of the normative character of law according to which it is not the obligation expression but rather existence of sanctions that determines whether a certain regulation is binding or not, it has to be stressed that this analysis examined respondents' views and their innate knowledge of a native language.

Differentiating features
For the purpose of this analysis the terms 'differentiating feature' or 'category' are adopted to refer to stylistic, denotational and contextual variations between obligation expressions.The term 'connotation' will be avoided here because it can be misleading due to its different meaning in the tradition of philosophy where it can also be regarded as a part of core meaning of an expression.The term 'variation' is used in line with the study by Edmonds and Hirst (2002).
The ten categories selected for analysis include: kind of obligation, modal strength, scope of obligation, sanctions, start or duration of obligation, deontic agent, register, contemporariness and origin.They will be now described in the following section.

Kind of obligation (recommendationorder)
This category describes whether the expression conveys an obligation or a recommendation.This means whether a statement is binding and categorical or rather gives conditional guidelines of behaviour.
Although this feature is one of ten that have been distinguished, it seems to be the only one that can be regarded as a denotational feature.This is because it conveys the most basic nuance and is therefore primary to other differentiating features.It is also mentioned in the dictionary definition which highlights its significance for the right discrimination between the expressions.This category is clearly related to the next one.

Modal strength (less obligatorymore obligatorynot obligatory)
It may be regarded a resultant of the above presented category and an existence of possible sanctions in a legal norm.This feature describes a level of what some researchers call 'modal' or 'deontic strength'.Within this category three options are distinguished: 'less obligatory', 'more obligatory' and 'not obligatory'.They express the strength of the obligation or in other words whether the language users perceive an expression as more or less obligatory in comparison to another in a certain context.

Scope of obligation (generalspecific)
The next category describes the character of the expression.A certain expression may be associated with a general or a more detailed obligation.For example, an expression may affect the statement in such a way that to some extent it is regarded vague, indefinite, wide in its scope which can be subject to interpretation and flexibility.
On the other hand, there is a subcategory 'specific' which means that a certain expression corresponds better with the clear and direct obligation, which is precise and does not leave any space for own interpretation or discretion.This unambiguity may result additionally from a clearly defined deadline or formulated detailed conditions under which some action should be performed.

Sanctions (no sanctionssanctions)
This feature outlines the potential existence of sanctions in case of acting against a certain norm.The choice of an obligation expression can be also connected with existing sanctions or lack of sanctions.This category is thus based on a privative opposition.

Start/ duration of obligation (starts right awaystarts in the future)
It defines whether the expression means that one has to take an urgent action determined in the norm or whether one can postpone some action according to the respondents.Within this category there are therefore two subcategories.In principle, the expression can to some extent affect the way the norm is comprehended.

Register (less officialmore official/ legalcolloquial)
This category defines style of the expression.It is related to the language register.Some expressions are more likely to be used in a formal or informal context.For example, some expressions are preferred in a legal register because of their official style, while others appear in a more colloquial context.'Legal' comprises thus the comments about festive or polite character of the expression or even about legalese.'Colloquial' means that expression is more typical for general language (yleiskieli) or even seems awkward when used in a legal text.Apart from two subcategories, i.e. legal and colloquial an additional subcategory 'neutral' is added.

Contemporariness (modernold-fashioned)
This category outlines the temporal dimension of the expression.Contemporariness describes whether it is perceived as an old-fashioned or modern expression.

Origin
It defines the etymology of the expression according to the respondents.It is divided into Finnish and Swedish, i.e. an expression of Finnish or a Swedish origin.

Deontic agent
This feature implies a tendency or a preference to use a certain obligation expression with a certain type of an agent upon whom an obligation is imposed, e.g. an institution such as a court or police.It does not concern a collocation but rather a case in which a certain expression is used with a particular agent because of the deontic strength of the expression.

Others
There is a separate category which contains a group of stylistic characteristics that did not match the distinguished ten differentiating features.It contains a number of respondents' single comments of particularly low frequency, which thus could not form a separate category on their own.The near-synonymy between obligation expressions can be confirmed by denotational and stylistic features like, respectively, being obligatory and legal genre (Table 5 and Table 6).

Results: comparison of the expressions
However, the profiles of their features differ especially in terms of scope of obligation, i.e. the most frequent feature for on -(t)tava was 'specific' which was indicated by 36.5% of respondents who ascribed at least one feature to on -(t)tava.This category can determine whether an authority has discretionary powers or not, e.g. according to a respondent, "Tulee ilmoittaa gives more discretion to a clerk (a policeman) than on ilmoitettava".Moreover, it may also be connected to the branch of law.It has to be stressed here that there is a special guideline for the legal drafters of the Finnish criminal law on using a certain obligation expression instead of an imperative form or passive, namely on -t(t)ava (on tuomittava, 'shall be sentenced'), as if it was regarded to express more explicit obligation (Lainkirjoittajan opas 12.9.6).Respondents' comments confirm that on -t(t)ava suits better a context which refers to a case that is more significant, like a crime to less significant, e.g. a petitionary matter.
Furthermore, the feature 'more obligatory' is associated more strongly with on -(t)tava as opposed to other deontic means (20.8 %), while for tulee only 2.7% of the respondents mentioned this feature.In contrast 12.3% of the respondents indicated 'less obligatory' for tulee and it can result from the fact that this expression is not perceived as fitting well with sanctions.Nobody named sanctions as an important context for tulee, but on the contrary, 3.4% of the respondents highlighted that it is used when rather no sanctions are conveyed.This can be in accordance with a subcategory 'recommendation' (14.4%) that was mentioned only for tulee and additionally, with a subcategory 'general'.The example is as follows: Example 1.
Todistuksen tulee sisältää osakkeiden merkinnän tai muun oikeuden käyttämisen ehdot.("The certificate shall indicate the terms of the subscription for shares or the exercise of the other right in question.")Respondent's comment: "These are some instructions for the drafter of the certificate."However, it seems that tulee in legal genre is an expression that usually imposes an obligation rather than recommendation.The semi-official translation of the Ministry of Justice of Finland can be very helpful in clarifying the kind of expressed obligation.The above mentioned example "The certificate shall indicate the terms of the subscription for shares or the exercise of the other right in question" or another one example: "Pelikasinotoiminnan harjoittaja ei saa päästää pelikasinoon henkilöä ja pelikasinotoiminnan harjoittajan tulee poistaa henkilö, joka on alle 18-vuotias (…)", which was translated into English as: "A casino operator shall not allow into the casino and shall remove from the casino any person who is under the age of 18 (…)" clearly prove that it is the obligation in question because the Finnish tulee was translated as shall which is an ordinary means of obligation in English legal texts.
According to the respondents, a certain expression can be regarded as an indicator that something has to be done immediately or can be postponed or done in an indeterminate future.This is encompassed by the category 'start/ duration of obligation': Example 2.
Kunnan tulee huolehtia alueellaan (…) ympäristön tilan seurannasta (…).The municipality has to take care of (…) environmental monitoring on its territory (…).Respondent's comment: "Tulee to some extent refers to the fact that taking care will begin in the future and that it will continue in the future."However, it is interesting whether associating the form tulee is not affected by meaning of the verb tulla 'to come' which conveys future in Finnish.Future can be expressed with a construction tulee + 3 rd infinitive, e.g.tulee tekemään, somebody 'will do' and it is adapted from the Swedish language (kommer att göra).Thus, it is interesting whether the deontic meaning has not been affected by the means of conveying the future.
The study also showed that the category of deontic agent is very hard to delineate and to confirm its real salience.Respondents' indication of this feature is not very frequent for both expressions (4.79% for tulee, 3.77% for on -t(t)ava), which shows that respondents are divided in their opinions and they do not show any clear evidence that this feature is connected significantly with the certain expression.Some argue that the obligation imposed on a court or any other representative of authority should not be expressed too strictly with an excessively categorical verb.Respondents state that "a court rules independently" and that "it is inappropriate to command the court in the wrong manner".As it is a court's task to apply the law and oblige other subjects to perform some actions, it is unnecessary to indicate it explicitly.The opposite views state that if an authority representative is the subject of an obligatory statement, "one can demand from him a more thorough knowledge of law" and the same applies to police which is "a representative of a legal institution" and the obligation can be more strictly expressed with a stronger expression.
Moreover, there were also single comments which suggested the importance of an opposition "institution -individual" that motivated the choice of an expression.According to this view, an obligation imposed on the authority, e.g. the judge, is expressed in a milder way than an obligation imposed on a citizen.This can be interpreted in such a way that in case court there is just a description of its ordinary tasks while in case of a citizen, it is a specific situation that is regulated.
The notion of a deontic agent playing a possible role in choosing deontic expressions was also mentioned by Attila (2016).In her study a certain obligation expression (on -(t)tava) was a clearly typical means of imposing obligation on subjects other than the authorities (Attila 2016: 12).This could confirm the results from this study.However, this topic should be studied more thoroughly to be able to draw any binding conclusions.Moreover, necessity in Finnish is also vastly expressed with indicative.With respect to this fact Attila suggests that the directive character of the expressions in propositions be more clearly expressed if they do not concern officials (Attila 2016: 23).
This possible opposition of the modal agents is related to the guidelines concerning the use of deontic means for Finnish translators of European Union acts.As far as the equivalents of shall are concerned, there is a note stating that the institutions of the European Union should be addressed with a verb in indicative mood, while agents other than European Union, such as member stateswith the necessity construction on -(t)tava.(Suomen kielen käyttöohjeita 2013: 60).
Respondents referring to features assigned to the residual category 'Others' were similarly frequent for each of the expressions (15.75% for tulee, 15.72% for on -(t)tava).Among single features in this category, an interesting example is a complement verb that follows the main verb (whether tulee + infinitive or on -(t)tava).Respondents chose the other expression from the list only because the other was considered stylistically inappropriate when conjugated in a particular grammatical form.This referred to a verb sisältää, 'to include' and evätä 'to refuse, deny, reject'.In this case respondents preferred to use the form tulee instead of conjugated forms sisällettävä and evättävä.
Although some stylistic categories like contemporariness and origin are not of central importance, they still convey some interesting additional meanings related to the use of analysed expressions.Respondents described a certain expression as being neutral and that is why it suited the legal context.'Neutral' can also be quite typical of standard language (kirjakieli), something between legal and colloquial, it is not apparently categorical.
As far as the 'origin' is concerned, some respondents decided not to choose tulee because of association with "bad, poor Finnish".It is thus very language-related, i.e.Finnish-related feature, nor is it mentioned often.However, it is interesting whether the deontic meaning has not been confused with the means of conveying the future that is expressed in Swedish in a similar way (as described earlier in the Results section).

Comparison between the groups of respondents
In this section the results of the survey are presented on the basis of educational background of the respondents.This can help clarify the possible differences in perception of the expressions between the groups.Below is a table that shows the internal diversity of the group of respondents (Table 7)."Yes" and "no" refer to the completed education in a certain arealawrelated or language-related.In order to show how completed education possibly affects the respondents' perception of the expressions, for the purpose of this analysis, three groups were chosen and contrasted: respondents with completed legal and language education, as well as respondents whose educational profile matches neither of the above areas (for the sake of clarity and ease of comprehension, hereinafter called 'lawyers' and 'linguists').Students form a separate category which has been excluded from the analysis, though.The same applies to the respondents who have completed both legal and linguistic education.However, only those who ascribed any feature to either of the expressions were included.The number of these respondents are shown in

Lawyers
Tables 9 and 10 show features that were indicated by respondents who declared completed law-related education.In this group for the verb tulee the feature "legal register" was the most common (29%), followed by "order" (21%) and by "others" (17%).The next three features are "recommendation", "neutral" and "specific".42% classified on -(t)tava as expressing "order" which makes it the most frequent characteristic.The feature "more obligatory" (12%) and a clear and specific character of the expression (36%) is also higher in the rating than for tulee which demonstrates that it clearly expresses obligation.Moreover, as far as the modal strength is concerned, "less obligatory" and "recommendation" have not been indicated by any respondent in this group.The differences in the features referring to modal strength for tulee are so slight that without detailed and formal tests it cannot be stated with certainty which of the subcategories (order or recommendation) prevails.
Both expressions are deemed similarly neutral, as it is their fifth most frequent feature in both groups.Other features may be regarded insignificant as they have been indicated by less than 10%, that is 1 or 2 respondents.

Linguists
Those who completed linguistic studies chose "legal" register to be the most typical feature of on -(t)tava (42%, Table 12).The next one is expressing "order" (37%) and "more obligatory" (24%) which shows apparently the unambiguously categorical character of this expression.This unambiguity is additionally reinforced by the chosen feature also with high frequency "specific" (31%).To the verb tulee (Table 11) linguists ascribed "legal" character (55%), followed by "recommendation" (21%) and "order" (16%).In comparison to the results for the expression on -(t)tava shown above, it is quite evident that linguists perceive the major difference between both expressions in terms of modal strength.

Respondents with neither law nor language-related backgrounds
Respondents with neither of the mentioned educational backgrounds constitute a group of 54.Among answers explaining the verb tulee the features legal register of the verb (36%), followed by order" (14% in this group also mentioned the category of "start of obligation" which was less popular in the ratings of other groups.The preference of the expression on -(t)tava is accounted for by its specific nature (40%) and its "more obligatory" character than tulee (32%).Moreover, 24% indicated it expresses order and 20% referred to its "legal" and official character.

Conclusions
The study shows that there are some features that can help differentiate between obligation expressions although they are nearsynonyms.Despite common features it can be seen that both expressions have different characteristics.The fact that some features were indicated either only for tulee or for on -(t)tava demonstrates that they are not fully interchangeable in all contexts.Their profiles differ most of all in terms of concentrations of features that denote the categorical character of an expression.
For this reason it can be stated that on -(t)tava could be more likely used in sentences where obligation is to be expressed in a more direct and clear manner.It also seems neutral, given its high frequency both in an official register in the corpus of legal texts, as well as in an everyday language.
In contrast, due to its low frequency in colloquial and spoken language, tulee is regarded more official and thus suiting the legal genre well.Sometimes it can be regarded as old-fashioned to some extent and that is why it might not be as universal as on -(t)tava.Respondents perceive it sometimes as less categorical than on -(t)tava, having a wider and a more indefinite meaning.Therefore, it may be preferred in contexts which should be left less detailed and precise, giving more space for interpretation (concerning time span, scope of obligation or modal agent).
In respect to the educational profile of the respondents it can be concluded that, surprisingly, all included groups of respondents refer to the modal strength in order to explain the difference between the expressions.The most common features ascribed to on -(t)tava in all groups were "order" and "specific" which shows that the expression is perceived clearly categorical.This matches also with the rating in the general results (Table 6).Features "legal" and "order" were among the top three for tulee in all groups, as well as in the top three in the general results (Table 5).Worth noticing is much more evident association of tulee with weaker obligation or recommendation in the group of linguists than amongst lawyers.This perhaps results from less experience with reading legal texts but may also be caused by the contextual factor which sometimes affects the way a certain expression is perceived.However, in order to be able to compare the groups more accurately and draw further conclusions, a more detailed statistical analysis is required.
Nevertheless, because both expressions are used in Finnish legal texts, although there are no official guidelines on their accurate usage, they should be considered equally expressing legal obligation.Semi-official ministerial translations confirm this interpretation and practice.

Table 1 .
Obligation expressions in the corpus of 50 most often used legislative texts in Finlex (Sept.2015). ):

Table 4 .
Included and excluded respondents.

Table 5 and
Table6show the distributions of frequencies of feature subcategories outlined by the respondents for tulee and on -(t)tava.Each respondent could indicate unlimited number of features characterizing given expression.Hence, the tables show both the number and proportion of respondents referring to particular subcategories.

Table 7 .
Educational profile of the respondents.

Table 8 .
Final number of respondents who were included into analysis.