Some Implications for Ērika: Implicatives in Danish, Finnish and Lithuanian

: The article deals with implicative verbs, i.e., verbs that, both in their affirmative and negative forms, carry implications as to the factual status of their propositional complements, e.g. manage, forget, bother etc. Karttunen (1971), who introduced the notion, already pointed out that a verb that is implicative in one language need not necessarily have implicative counterparts in other languages. It is conceivable that some languages have semantic groups of implicatives not represented, or less well represented, in other languages, and this deserves to be investigated. In this article the authors offer just a very preliminary exploration based on three languages, one North Germanic, one Fennic, and one Baltic. They show that even such a small sample may reveal interesting differences. The authors also pause over certain general tendencies in the semantic development of implicatives. While most of the work on implicatives has been done in the tradition of formal semantics, the authors show that a more cognitively oriented approach (invoking mechanisms of subjectification) can yield valuable insights into the polysemy of implicatives.


Introduction1
In 1971 Lauri Karttunen formulated the notion of implicative verbs, by which he means verbs that, both in their affirmative and their negative varieties, carry certain implications as to the factuality of the situation described by their propositional complement. A typical example is manage2: (1) Solomon managed to build the Temple.
(2) Solomon didn't manage to build the Temple.
(1) entails Solomon built the Temple, whereas (2) entails Solomon did not build the Temple. With some verbs polarity is reversed, but a twoway implication still holds: (3) Solomon forgot to build the Temple.
(4) Solomon did not forget to build the Temple.
(3) entails Solomon did not build the Temple, whereas (4) usually3 entails Solomon built the Temple. Karttunen's article came a year after Kiparsky and Kiparsky's (1970) equally famous article on factivity. The latter gave rise to the notion of factive complement-taking verbs, which presuppose the factual status of their complements, cf.
(5) Solomon regretted having built the Temple. (6) Solomon did not regret having built the Temple.
As both (5) and (6) can only be felicitously used if the Temple has indeed been built, the building of the Temple is, in this case, a presupposition. Karttunen points out that implicative predicates also carry certain presuppositions, e.g. (1) and (2) usually presuppose that there were serious obstacles to the building of the Temple, and that Solomon made a directed effort at overcoming these obstacles.
Karttunen furthermore points out that a verb that is implicative in one language need not have this property in other languages. He gives a list of Finnish implicatives that have no English equivalents. An example would be arvata 'guess': (7) Finnish Arvasimme tulla juhlaan oikeaan aikaan. guess.pst.1pl come.inf party.ill.sg right.ill.sg time.ill.sg 'We contrived to arrive at the party at the right time.' (i.e. 'we guessed what the best time to arrive would be and arrived at that time ') Implicatives have hitherto drawn attention mainly from scholars working in the tradition of formal semantics. Their purpose has been to gain insight in the exact nature of the presuppositions associated with the use of an implicative, to establish what exactly is asserted by an implicative verb, etc. But what we know about the existence of language-specific (types of) implicative verbs suggests it might also be worthwhile looking at implicatives cross-linguistically. Noonan (2007,139) briefly mentions implicatives (under the name 'achievement predicates') in his overview of complementation types but says nothing about typological variation, probably because no interesting variation in expression specific to implicative verbs is to be expected: whatever pattern is used in a given language for complement types involving systematic identity of main-clause and complement-clause subjects, such as those of desiderative or modal verbs, will probably also be used with implicative verbs. These will belong to what has been called the 'stateof-affairs' type of complement clauses, as opposed to the propositional type used, e.g., with verbs of saying, knowing etc. (cf. Kehayov & Boye, 2016, 812-818). Differences will concern the semantic domains in which implicative verbs occur; they are therefore a subject for lexical typology.
This article offers no more than a very preliminary exploration of the cross-linguistic study of implicatives; it is based on a small conveniency sample comprising the languages on which the authors happen to be working. Still, the authors hope that this exploration, as well as their thoughts on the semantic description of implicatives, will prove to be of some use. For considerations of space, we basically leave phrasal implicatives (Karttunen, 2012) out of consideration.
The structure of the article is as follows. In section 2, we will present some general considerations on the semantic properties of implicatives and the different ways in which one can look at them. Next, we will attempt a rough and preliminary semantic grouping of implicatives and try to identify a few groups where the languages under discussion show interesting differences. By way of conclusion, we offer a few preliminary generalizations.

On the semantics of implicatives
Almost all that has been written on implicatives has been written in the tradition of formal semantics. Authors working in this tradition have raised important questions and gained valuable insights, and we will briefly discuss a number of interesting points touched upon in the literature. We will argue, however, that a discussion of the semantics of implicatives can also profit from a more cognitively oriented approach.
Karttunen characterizes the semantics of constructions with implicatives by pointing out the following features: (i) there are certain presuppositions associated with the use of implicatives, and (ii) a presupposed state of affairs is viewed as a necessary and sufficient condition for the realization of what is expressed by the complement. Consider (8): This sentence carries background presuppositions to the effect that John is under an obligation to turn off the light (e.g., before leaving his workplace), and that he is willing to carry out this obligation; there is also a specific presupposition to the effect that John's remembering the obligation is a necessary and sufficient condition for his performing it. The act of uttering (8) amounts to stating that this specific presupposition is satisfied, and the truth of the embedded proposition automatically follows. A crucial element of Karttunen's analysis is that, truth-conditionally, the implicative predicate adds nothing to what is asserted by the embedded clause; all it adds is the commitment to certain presuppositions. In a later formulation (Karttunen, 2014), the notion of presupposition is replaced with that of conventional implicature. This lack of truth-conditional content -the actual occurrence of the event described in the embedded proposition being automatically captured by the status of 'sufficient condition' -is one of the elements in Karttunen's analysis that have been challenged. Writing on manage, Baglini and Francez (2016, 546) claim that this verb "makes a non-trivial truth-conditional contribution". A sentence of the type 'manage p' "presupposes the familiarity of a 'catalyst', a causally necessary, but causally insufficient condition for the truth of p, and asserts that the catalyst actually caused the truth of p". This proposal for improvement of Karttunen's analysis seems intuitively convincing because, assuming free will in humans, no external circumstance or catalyst can ever be a sufficient condition for the accomplishment of an act.
But there are problems of a more general nature with the reliance on presuppositions in describing the semantics of implicatives. Coleman (1975) draws attention to 'shifting' presuppositions with manage. Karttunen assumes 'trying' is an essential presupposition for the use of manage, but as Coleman points out, there are many contexts where this does not apply: (9) My dog manages to get clawed by every cat that comes along.
Moreover, the subject of manage may be inanimate or ambient, in which case no attempts could possibly be involved: Coleman suggests this can be accounted for by operating with a hierarchical ordering of presuppositions, presuppositional elements incompatible with the context being successively filtered out: if the presupposition 'the subject tried to achieve p' fails, the weaker presupposition 'p is difficult to achieve' is substituted, and if that fails as well, what remains is 'p is unlikely' .
This account in terms of 'vanishing presuppositions' was devised in order to avoid recognizing polysemy (or 'homonymy' , as Coleman, true to the spirit of the times, puts it). The monosemy assumption pervades the whole literature in the tradition of formal semantics, and therefore also almost all of what has been written on implicatives. The assumption is usually that a linguistic sign can only have one meaning, and that apparent meaning differences must originate in pragmatics. Perhaps the study of implicatives could profit from an approach in the spirit of cognitive linguistics, based on the assumption that one linguistic sign may correspond to several conceptual structures mutually linked by recurrent patterns of semantic change such as metaphor, pragmatic strengthening etc. and thus forming a network, as has been argued in a large body of literature starting with Brugman & Lakoff (1988). This principled polysemy approach is also more in keeping with the purposes of cross-linguistic comparison, which is badly served by a monosemic approach.
A use like (9) could be characterized as an instance of irony, in which case we would have to look for an explanation in pragmatic terms. But we could also consider a semantic account in terms of subjectification, that is, a shift "from 'propositional' to 'textual' meanings, or from describing an external Situation to reflecting evaluative, perceptual, or cognitive aspects of the 'internal Situation'" (Langacker, 1990, 16, with reference to the pioneering work of Elisabeth Closs Traugott, starting with Traugott, 1982). When we compare (1) to (9), we may note a shift from an obstacle in the real world (the material difficulties standing in the way of the erection of the Temple) to an obstacle in the world of beliefs -the building of the Temple being a formidable task, its successful completion is unexpected, and unexpectedness is what remains from the original meaning in (9). A further metaphoric transfer enables the shift to inanimate or ambient subjects in cases like (10). Such processes are semantic, not pragmatic in nature, and they form regular patterns: as we will see below, a similar shift from reference to states of affairs in the extralinguistic world to evaluative meanings is characteristic of many implicatives. To discover a general pattern of this kind seems somehow more profitable than seeking ad hoc reasons for shifting presuppositions in every particular case.
In a more recent publication, Karttunen (2014) gives as a rule of thumb for describing the meanings of implicatives that they all "suggest […] or conventionally implicate that there is some obstacle that must be overcome for the infinitival clause to be true". We agree with this characterization but prefer to assume that the obstacle to be overcome is not conventionally implicated but linguistically encoded. As a result of different semantic processes the 'obstacle' may be variously reinterpreted, as the example of manage shows. But the obstacle may also, as long as it is situated in the extralinguistic world, be of different kinds, and it is here that cross-linguistic variation may manifest itself. Karttunen has already drawn attention to Finnish as a language rich in implicatives, adding jocularly that "having a specific implicative verb available for so many obstacles makes Finnish a great language for pithy excuses" (Karttunen, 2014). If this is the case, then perhaps Finnish implicatives are especially well represented in certain domains, whereas other languages specialize in other domains? This is, in our view, an interesting research question. It requires, of course, an adequate classification, which it will perhaps be possible to give only at a later stage, on the basis of broader empirical data. The classification used here is only preliminary and largely intuitive. The obstacles Karttunen refers to may remain unspecified, as in the case of manage, or they may be specified, as in the case of remember. In this classification, we set apart a group of implicatives that are non-specific in the sense of being noncommittal as to the nature of the obstacles, and several groups of more specific implicatives singled out according to the nature of the obstacles involved -mental, emotive, social, physical and spatio-motoric.
3. Non-specific implicatives All three languages examined here have several non-specific implicatives, by which we mean implicatives referring to an agent bringing about the occurrence of an event in spite of certain unspecified obstacles, such as manage and succeed. We can distinguish two subtypes here, one laying more emphasis on the subject's skills whereas the other emphasizes external circumstances. The second type has a preference, in many languages, for impersonal constructions, with the subject-agent as a datival argument. This can be seen in Danish, which has evne 'manage' , representing the first subtype (14) and lykkes 'succeed', representing the second (15) A second group puts less emphasis on the adverse influence of circumstances and on the need for a subject to surmount these. The occurrence of the event is driven by circumstances rather than by the subject's will: (20) Claudia became the project scientist of the Rosetta mission where she got to work with NASA's European counterpart, the European Space Agency.6 Implicatives of this type seem to develop out of acquisitive verbs like 'get' (ultimately also a source for modal verbs, cf. van der Auwera, Kehayov & Vittrant, 2009) and verbs of motion like 'come', the borderline between the two not always being clear (e.g. English get can be acquisitive but is also used as a motion verb: How did you get there?). Danish has both sources. The productive construction has 'come': (21) Danish Hun kom bestemt ikke til at kede sig. she come.pst certainly not till to bore.inf refl 'She definitely did not have occasion to be bored.' (KorpusDK) Danish få 'get' is, at least in the contemporary language, severely restricted lexically, occurring mainly with a few verbs of cognition and perception in what appear to be fixed expressions: Here the subject is a theme in motion towards a certain type of situation, but the spatial conceptualization may also be reversed, with the event as a theme and the agent as a recipient/goal, a pattern to be found in Lithuanian: A reverse non-specific implicative is 'omit', here represented only by Danish undlade, without exact equivalents in Lithuanian and Finnish though phrasal implicatives may be used to fill the gap. It is interesting in that this verb meets the formal conditions for implicatives but presuppositions differ considerably according to polarity. The affirmative form suggests failure to comply with an obligation or general norm of behaviour, but remains vague about the reasons of non-compliance (forgetfulness, carelessness, willful non-compliance etc.). The negative construction is usually ironical and suggests a person indulging a none too creditable character trait: Discreditable character traits are not proscribed by formal laws but may invite censure, so that, in order to offer a unified account for (24) and (25), we could suggest that the subject in (25) fails to comply with an unwritten law requiring people to refrain from indulging bad tendencies. The actual presuppositions behind (25) are different, and the shift seems to be, at first glance, pragmatic, but it might in fact reflect a general tendency for implicatives to acquire an evaluating function, on which see below.

The mental sphere
All three languages have specific implicatives singling out a mental process as determining the event named in the embedded predicate. The most widespread mental implicatives include 'remember (to do something)' and 'forget' . Another aspect of mental activity is 'presence of mind' , i.e. the ability quickly to devise an adequate course of action apposite to the situation. Finnish uses its verb 'know' for this meaning: Lithuanian and Finnish use the verb 'understand' in more or less the same sense (in Lithuanian a special reflexive form is used): Other expressions, like Danish falde ind and the Lithuanian phrasal implicative (cf. Karttunen, 2012) šauti į galvą, use a motion metaphor and a syntactic structure in which the thought is the subject-theme and the person is an indirect object: Some languages have implicative verbs whose use presupposes that, all other conditions being satisfied, the accomplishment of an action depends only on an act of volition on the part of the subject. English bother belongs to this type. Finnish, true to its reputation as a language rich in implicatives, has several verbs to convey this meaning: viitsiä, välittää (only in negative sentences) and (apparently an innovation under English influence) vaivautua: (31) Finnish En viitsinyt ottaa mitään riskiä. neg.1sg bother.part.pst take.inf any risk.prt.sg 'I did not bother to take any risk.'13 Interestingly, Lithuanian's sister language Latvian has borrowed the Fennic verb reflected in Finnish viitsiä (Latvian nevīžot, used only with negation). Danish seems to have no exact equivalent: gide is usually given as a counterpart of bother in dictionaries, but it does not behave as an implicative. In Lithuanian the equivalent of bother would be pasivarginti 'take the trouble' (usually in an ironical sense): (32) Lithuanian [Pirmiausia stebina tai, kad korespondentė, ketindama rašyti straipsnį,] net ne-pasivargino pasikalbėti ar prisistatyti. even neg-take.the.trouble.pst.3sg talk.inf or introduce.oneself.inf [What is most surprising is that the correspondent, intending to write an article,] didn't even bother to talk [to us] or introduce herself.'14 5. The emotive sphere Also well represented in the three languages, and probably rather universal, is the emotive sphere. The implicative refers to a state of mind viewed as a necessary condition for the realization of an event. The obstacle the subject has to overcome is a mental inhibitor such as shame, compassion, disgust etc. As such feelings are experienced in relation to other persons, socio-emotive would perhaps be a more appropriate term, but we will reserve 'social' for those meanings specifically connected with social hierarchies.
A typical example of a socio-emotive implicative would be Danish naenne, which can be translated as 'to bring oneself to do sth' . The obstacle is the subject's sense of delicacy: (33) Danish Men han naennede alligevel ikke at vaekke hende. but he bring.oneself.pst nevertheless neg to wake.inf her 'Nevertheless he couldn't bring himself to wake her.' (KorpusDK) Finnish has a whole set of reverse implicatives with the meaning 'be prevented from sth by shyness': kainostella, arastella and ujostella. They are presumably differentiated by subtle shades of meaning.
(34) Finnish Todellisista tunteistaan hän-kin arasteli puhua. true.elat.pl feeling.elat.pl.3sg 3sg-also feel.shy.pst.3sg talk.inf 'He/she was also shy to talk about his/her true feelings.'15 Not less abundantly represented is the opposite type, describing a situation in which a person surmounts inhibitors of various kinds. This type shows very fine distinctions as to the exact nature of the inhibitor: we have juljeta 'have the impudence', kehdeta 'not be constrained by delicacy, shame etc. ', iljetä 'surmount one's repugnance', sumeilla 'have scruples' (only with negation) etc. Though the King had, perhaps, to surmount a natural social barrier that would have prevented him from meeting the petitioners, he may not have given much thought to this aspect of his decision and what the implicative verb adds is a certain evaluation to the effect that King's readiness to meet the petitioners was extraordinarily gracious. The original notion of a social obstacle hardly ever occurs here. The 'condescend' type is represented in all three languages dealt with here, though, in our democratic times, it is probably restricted to ironical use: (39) Danish Peter nedlod sig ikke til at svare. Peter lower.pst refl neg till to answer.inf 'Peter did not condescend to answer.' (KorpusDK) In addition to malonėti and teiktis 'deign, condescend', Lithuanian also has the reverse implicative pasididžiuoti 'refrain from doing sth out of pride' (this verb also means 'feel or express pride'), whose negative form nepasididžiuoti can be translated as 'condescend': This seems to imply an evaluation that is not merely social but has a moral dimension, but anyway what the verb adds is only an evaluationwe do not even know whether the subject has to surmount any scruples. In this sense a verb like alentua is similar to implicatives in the social sphere: they do not add anything to the simple verb truth-conditionally, but add an evaluation in terms of societal norms.

The sphere of physical sensations
Can the obstacle to be surmounted, as represented by an implicative verb, be a physiological sensation rather than an emotion? It can: this sphere is represented by Finnish tarjeta, which reflects sensitivity to temperature: Here most languages would have a less specific verb like 'dare' , but the Finnish verb additionally specifies that the subject must brave the cold in order to accomplish the feat described by the embedded infinitive. The verb tarjeta appears to occupy an isolated position among Finnish implicatives; there is no analogous verb referring to the act of braving the rain or ignoring a headache. So, while this example illustrates the general tendency of Finnish to be quite specific in its characterization of obstacles to be surmounted, the physiological sphere is weakly represented in comparison to the socio-emotional one. It remains to be investigated whether equally specific implicatives in the physiological sphere can be found in other languages.
8. The spatio-motoric sphere Lithuanian has not been mentioned until now as a language particularly rich in specific implicatives, but in fact it has one interesting group of these, viz. a small group of verbs describing the ability to accomplish some action as determined by the ability to overcome certain spatio-motoric limitations, partly determined by bodily predispositions. Pasiekti 'reach' is one of them: This use of pataikyti is ironical, of course, as the implication is the people referred to didn't bother to throw their empty cans into a dustbin. But more interesting here is the extended, metaphorical use of pataikyti, as in: 'It [sc. the homework] helps me to assess whether a child just hit upon the correct answer when asked or really understood the content of the lesson.'25 Here pataikyti means 'by chance more than understanding, find the right way to do something' . Interesting questions are raised by the negated variety nepataikė atsakyti, which could be interpreted as 'answered but not to the point' . This would suggest pataikyti is, in this use, no longer implicative but factive. The interpretation of such sentences is not straightforward but such a shift to factive status would not be quite unexpected: as the implicative's meaning becomes evaluative, the content of the complement clause may cease to be interpreted as being conditional on what is expressed by the complement-taking verb and become a presupposed event subject to evaluation.
A metaphorical extension of the kind found with pataikyti is also found with the verb aprėpti, originally 'encompass (with the arms)', which belongs to the same group as pasiekti or pratilpti, and indeed also behaves as an implicative complement-taking verb, but seems to be attested only in an extended, metaphorical sense, as 'manage to cope with a large number of tasks': Interestingly, implicatives in the spatio-motoric domain are also found in Finnish, e.g., ulottua 'reach, manage to reach' (a counterpart to Lith. pasiekti), and mahtua 'fit into sth, find place enough to do sth' (a counterpart to Lith. tilpti):  (RNC, from Fedor Knorre, 1973) However, speakers of standard Russian do not readily accept such constructions, and constructions like (53) could be an occasional phenomenon.
Verbs belonging to the group discussed here show a tendency to select clausal complements in other languages as well, but not necessarily with the verb in its original spatio-motoric meaning. The Danish verb nå 'reach' is also an implicative, but its spatial meaning has evolved into a general meaning of ability and ultimately into the temporal meaning of 'manage to do something within a certain time frame': The shift to more general and abstract meaning can be compared to that observed in Lith. pataikyti, discussed above, which means 'manage to do something in the way required by convention, at the right moment etc. ' , or in Lith. aprėpti, which has acquired a quantitative meaning. Such extended and metaphorical uses appear regularly, but the specific feature of Finnish and Lithuanian is that they have a small group of spatio-motoric verbs that, as implicative predicates, have remained stable at the spatio-motoric stage instead of moving on to more abstract meanings, although these also arise and co-exist with the more concrete ones. The Finnish-Lithuanian convergence (perhaps extending to Slavic) is interesting, though it is not clear whether we are dealing with an areal feature.

Time frames as obstacles
Several languages have specialized implicative verbs describing a situation in which a person succeeds in performing a task in spite of being pressed for time. Danish nå has already been mentioned. Lithuanian has a verb of the same type: Though we can describe 'lack of time' as an obstacle to be surmounted (analogous to lack of space in the case of tilpti above), it seems verbs of the type dealt with in this section are prone to extended uses in which the implicative verb ceases to give an internal characterization of the situation and shifts to temporal location of the event with regard to other events, thus performing a function close to that of phasal adverbs. Cf. (58) Finally, parallels are also found in Slavic (cf. Russian uspet', Polish zdążyć), both in the original sense of 'obstacle' and in the temporal-location sense. Could this be an areal feature?
Only Lithuanian seems to have a reverse implicative counterpart pavėluoti, meaning 'be late': As Karttunen notes, implicatives usually imply that a certain obstacle has to be surmounted in order for the subject to bring about a certain event. The presence of obstacles to be surmounted as a presupposition associated with the use of implicative verbs is a convenient point of departure for a general characterization of implicatives, but certain groups of implicatives are prone to shifts of different types according to the semantic group to which the implicative belongs. In a general way, we can describe these shifts as instances of subjectification. As what should count as an obstacle is often a matter of subjective evaluation, the path to subjectification is open to virtually all implicatives, though in some cases the subjectified use is felt to be ironical, as in He managed to offend half of the electorate, etc. Whereas for verbs like 'manage' in their most basic sense it is an object of controversy whether the implicative makes a truth-conditional contribution or not, it is clear that there is no such contribution in the case of subjectified implicatives: here the sole difference is in the speaker's evaluation.
The subjectified reading of implicatives consists, in most cases, in an evaluation of a person's behaviour in terms of social conventions, ethical standards etc. In some cases, the evaluative element is lacking: verbs like Finnish ehtiä, Lithuanian spėti may simply express relative location in a time-scale, while Lithuanian aprėpti in (50) has shifted to a quantitative characterization. But in such cases as well, a process of subjectification is at work.
Our exploration confirms that languages may differ with regard to the degree of differentiation and specificity of meanings in the implicative domain. The non-specific implicatives may show occasional gaps in individual languages, but these are not of broader interest. More general tendencies manifest themselves where a language has a semantic class of specific implicatives lacking in other languages. Both Lithuanian and Finnish show instances of this. The spatio-motoric group looks like an interesting Finnish-Lithuanian convergence -could it be an areal feature? Finnish tarjeta 'feel warm enough (to do something)' points to the existence of a 'physiological' group, though we have as yet found no more examples. The Finnish-Lithuanian convergence with regard to the spatio-motoric type suggests that areal patterns might perhaps be discerned. The present cross-linguistic study was highly selective, the choice of languages being determined by the authors' linguistic expertise. A broader investigation on implicatives in European languages would no doubt yield more trustworthy and revealing results. But we