POLISH IN THE LIGHT OF GRAMMATICALIZATION THEORY

The paper is concerned with grammaticalization, a type of language change whereby lexical items, in specific contexts, come to serve grammatical functions, and grammatical items acquire new grammatical functions. The aim is twofold: to shed light at the main properties of grammaticalization, and to demonstrate its applicability to Polish data. Some prominent examples in Polish are discussed: the grammaticalization of modals, imperative and avertive constructions. The paper closes with a non-exhaustive list of leads for further research into grammaticalization in Polish.


Introducing 'Grammaticalization theory'
The present article aims at giving an overview of recent developments of the theoretical frameworks subsumed under the heading 'Grammaticalization Theory' on the one hand, and at showing ways of possible applications of this framework to the analysis of both synchronic and diachronic data from Polish on the other.The term 'grammaticalization', sometimes also in the form 'grammaticisation' (henceforth gzn), was originally coined by Antoine Meillet in 1912, but has become popular in linguistics only since the seminal work by the German linguist Christian Lehmann 'Thoughts on Grammaticalization ' first published in 1982, republished and expanded 1995 and 2002.Under the term gzn we understand a type of language change whereby lexical items and constructions, in specific contexts, come to be used as grammatical, or whereby grammatical items acquire new grammatical functions (Hopper and Traugott 2003, p. 1).It turns out that these unidirectional changes are not language specific, but are governed by cross-linguistically or even universally valid principles, which is to be explained by universals of human cognition.Gzn theory provides tools to discover how different grammatical elements are related to each other, how systematic polysemy can arise and why certain linguistic forms have simultaneous lexical and grammatical functions (Heine and Kuteva 2002, book cover).It is not an exaggeration to say that gzn theory belongs to one of the most thriving fields within European Diachronic Linguistics and has gained considerable influence on linguistic theorizing and the development of linguistic thought not only among typologically oriented scholars, but also to an increasing degree among scholars working on individual languages or language families.An important forum bringing together linguists working on different languages is the triennial conference "New Reflections on Grammaticalization" held in Potsdam (1999), Amsterdam (2002)2 , Santiago de Compostela (2005) 3 and Leuven (2008).These conferences and their proceedings illustrate that gzn theory has become a viable international platform for the exchange of thought between linguists and has thus contributed to overcome the boundaries of the individual national philological traditions which used to hinder the exchange of thought between linguists from different countries.
An important characteristic of gzn theory is its holistic understanding of language, i.e. it rejects the idea that language should consist of closed boxes (modular view) and, instead, integrates structural aspects of phonology, morphology, and syntax with semantics.A second important facet of the holistic view is the abandonment of the structuralistic dichotomy between diachrony and synchrony in the sense of Ferdinand de Saussure: grammaticalization acknowledges the importance of diachrony as an explanatory factor in synchronic linguistics (see Hopper and Traugott 2003;Kuteva 2001).In other words, grammaticalization makes recourse to the panchronic approach under which language is viewed as a dynamic system, and diachrony is hypothesized as manifesting itself in synchrony.
Since the object of study of gzn researchers is the transition of lexical elements into grammatical markers it comes as no surprise that nearly all scholars agree that there cannot be a strict distinction between the two linguistic domains lexicon vs grammar.Linguistic categories are generally assumed to have fuzzy boundaries and to contain heterogeneous elements which is in line with Prague Functional School which propagated the idea that linguistic categories "should not be regarded as 'boxes' with clear-cut boundaries but as formations with a compact core (centre) and with gradual transition into a diffuse periphery which, again, gradually passes (infiltrates) into the peripheral domain of the next category."(Daneš 1966, p. 11). 4  Gzn theory also shows a certain affinity with Indo-European studies and etymological research, mainly in the field of the development of grammatical markers.Thus, works like the 'Etymologický slovník slovanských jazyků.Slova gramatická a zájmena' (Havránek 1980) or the etymological dictionaries of Polish (Brückner 1927, Bańkowski 2000) offer rich sources of material for potential gzn research.
Turning now to the topic of the present article, one has to state that gzn theory has not yet made its way into Polish linguistics.The aim of the present paper, thus, is to show the advantages and strengths of this framework and to develop a kind of research agenda for the study of grammaticalization phenomena in Polish.We are not yet in a position to present wide ranging results of gzn research on Polish and, therefore have to restrict ourselves to a small number of case studies and a sketchy outline of phenomena usually treated as instances of gzn processes.

The state of the art
Defined as a gradual development of grammatical words from content words (Meillet 1912), gzn provides an account for how grammar comes into being (cf . Hopper 1991;see Bybee 2003a).It shows that grammatical items, albeit abstract, should not be viewed as arbitrary, or non-motivated elements: instead of emerging ex nihilo, grammatical items can actually often be traced back to words with concrete meanings.The well-known development of the English future marker from the English movement verb to go will serve here as an example.Thus, the progressive form of the verb to go once exhibited only spatial directional use, as in They are going to Windsor (see Bybee 2003a, p. 150).In constructions in which going to came to be used with a verb, as in They are going to Windsor to see the King, a strong notion of intention was present, along with the primarily spatial meaning (Why are they going to Windsor?To see the King.).When the intention meaning prevailed, it was possible to infer future actions as in He is going to buy a house, which can be understood either as an intention or a prediction about future actions.Gradually, over a span of several centuries, going to came to be used with unambiguously future readings, with no meaning of intention or movement implied (That tree is going to lose its leaves.), a use that did not exist in Shakespeare's time (Bybee 2003a, p. 150).
Today, gzn is seen as "arguably the most widely discussed type of linguistic change" (Fischer and Rosenbach 2000, p. 1), and this reputation is believed to be due to the following properties.First, cross-linguistically identified instances of gzn tend to show remarkable similarities (see e.g.Bybee et al. 1994;Lamiroy 2003).Thus, it is not only the case that future markers developed out of the volitive verbs in areally and genetically unrelated languages such as Swahili (Heine 2003), English (Bybee et al. 1994;Traugott 2003), Serbian/Croatian and Mandarin (see Bybee and Dahl 1989), but their semantic development appears to have gone through the same stages (Bybee et al. 1994): volitional verb > intention > prediction > future Second, gzn tends to proceed in one direction, viz.from concrete to abstract but not vice versa (e.g.Haspelmath 1999;Kortmann 1999).The question as to whether gzn is unidirectional ignited much debate (see Haspelmath 1999;Fischer et al. 2004).Although some instances of change in the opposite direction, viz.from abstract to concrete have been identified (see Ramat 1992), such cases are far fewer than the numerous examples confirming the unidirectionality hypothesis (Newmeyer 1998, p. 275-6;Haspelmath 1999).Many instances of change from abstract to concrete, initially believed to be counterexamples to unidirectionality of grammaticalization, were later proved instances of lexicalization.For instance, the use of if and but in ifs and buts cannot be treated as an instance of reversed grammaticalization (from connector to noun) given that they are taken out of their construction and used metalinguistically (Haspelmath 1999(Haspelmath , p. 1064)).A similar example of lexicalization, that may be mistaken for an instance of degrammaticalization, can be found in Polish in the expression Każde dlaczego ma swoje dlatego.There are no studies available on cases of assumed degrammaticalization in Slavonic, but we would argue that such an example could be the Slovak modal verb musiet' 'must' which in negated contexts in colloquial speech acquired the new lexical, i.e. non-modal meaning 'to like' as illustrated in the following example from Findra (2008, p. 10):
'Are you going to the Karel Gott concert?I can't stand Gott.' Examples of this kind have initiated a discussion on the differences between gzn and lexicalization.As a result many interesting studies appeared, such as Brinton andTraugott (2005), andHimmelmann (2004).More importantly, however, no instances of complete reversals of gzn have been identified so far (Newmeyer 1998, p. 263).
Third, gzn is the cross-componential change par excellence: it may, but does not have to, affect all levels of language structure, viz.phonology, morphology, syntax, semantics and pragmatics (cf. McMahon 1994, p. 161).This can be illustrated by the Old Polish verb niechać which originally functioned as a lexical verb 'to leave' as in (2) and later developed into an imperative marker as in (3) and ( 4).
Although the effects of gzn may be visible at different levels of language structure, many scholars consider gzn to be first and foremost semantic in nature (cf.Bybee and Pagliuca 1985;Hopper and Traugott 2003, p. 76).In other words, semantic change from concrete to abstract (or abstract to even more abstract) is believed to be the precursor of structural change and is, as such, considered to be essential to any instance of gzn (see Bybee et al. 1994, p. 37).The changes in form that take place in gzn come as a result of the process of automatization of frequently occurring strings of linguistic elements (Bybee 2003a, p. 153).Grammatical meaning is, thus, established, long before the form becomes affected, which implies that formal changes (e.g.erosion, fusion) are reserved for later stages of grammaticalization (cf.Heine 2003;Hopper and Traugott 2003, p. 157-158).
This brings us to the subject of parameters of gzn, i.e. changes, either semantic or formal, that an element may undergo while grammaticalizing.5 Grammaticalisation comes in degrees and is understood as a complex macroprocess which is grasped in terms of different aspects located at the paradigmatic and the syntagmatic level.Closely related to the semantic-categorial shift is decategorialization, which is an effect of, and in that sense concomittants of, the semantic-categorial shift, and thus co-constitute an item's decategorialization (see Hopper 1991;Heine 2003;Drobnjakovic 2009).The Polish imperative marker niech can again serve as an example: it is indicative of the semantic shift 'to leave' > imperative and the categorial shift lexical verb > analytical imperative marker.The decategorialization mechanisms of imperative niech are actually the effects that this semantic-categorial shift has on the linguistic behaviour of the gram: for instance, niech can no longer be complemented by noun phrases.Decategorialization, thus, means that a grammaticalized item has lost syntactic properties of the category it belonged to before it embarked upon gzn.The decategorialization of niech is reflected in i) paradigmaticization (the integration into the imperative paradigm, see Lehmann 2002), ii) a change regarding complementation pattern (change from noun phrase to whole clause, see e.g.Bybee 2003a), iii) obligatorification (after an item starts grammaticalizing and becomes a part of a new, grammatical category, its interchangeability decreases; consequently, a language user has much less choice in replacing a grammatical than a lexical item with a synonym), and iv) fixation (the position in the clause becomes fixed).Lastly, apart from the main mechanisms, there are a number of gzn components which occur only at a later stage can be considered as secondary.The secondary parameters are such that they may well be present, but their absence would not jeopardize an item's gzn status: they rather proceed from an increase in frequency and the phenomenon of automatization (see e.g.Bybee 2003a).This group of gzn mechanisms consists of cliticization and affixation (both instances of coalescence) and phonological erosion on the one hand, and the change related to the semantics of the elements filling the subject and complement slots within the argument structure of a gram on the other.Cliticization, affixation and erosion are different outcomes of the process of automatization and repackaging (see Haiman 1994;Bybee 2003b).With repetition, sequences of elements that were previously independent start being used as a single unit or a chunk (Bybee 2003a, p. 153).As a result, the elements involved are no longer perceived individually and the whole chunk starts to undergo a reduction in form.This change can be illustrated by the development of the Polish past tense which goes back to an analytical construction consisting of the copular być plus the l-participle as attested in documents of the 14 th and 15 th centuries: przysiągł jeśm > przysiągł-em.Here the auxiliary eroded and developed into a clitic occupying a fixed position in the clause.
As far as the changes affecting the semantics of the elements filling subject and complement slots of a gram are concerned the range of items that can occupy subject and complement positions of a grammaticalized item becomes broader.Thus, whereas in Old Polish the modal powinien in its modal meaning was restricted to human subjects, in later periods it became compatible with inanimate subjects as illustrated in the utterance by Mickiewicz (see Hansen 2001, p. 331 f): (5) Epopeia powinna mieć akcyę ważną.(Uwagi nad Jagellonidą 1818) The main reasoning behind a division of the gzn mechanisms presented above comes from what has been repeatedly mentioned in research on gzn, viz.that not all mechanisms have to be involved for a grammaticalized element to be acknowledged as such (see Hopper and Traugott 2003).Gzn proceeds in stages and each of the stages is associated with specific mechanisms (see e.g.Heine 1993;Hopper and Traugott 2003).Given that gzn does not have to proceed all the way, i.e. the development may be 'frozen' at some point (see e.g.Hopper and Traugott 2003;Hansen 2004a), some mechanisms may not at all be involved.

Gzn of constructions (cxn)
In the last couple of years we witness a shift away from a so-called 'morphemecentric' view of gzn phenomena towards a more constructional view.Therefore we see attempts to integrate more formal syntactic aspects of linguistic structures like dependencies, phrase structure, word order, control etc. into research on gzn phenomena.This has lead to a considerable convergence with the syntactic theory of 'Construction Grammar' (see Fried and Östmann 2004;Traugott 2003;Noël 2007).The point of departure is the observation that one and the same lexical item can develop in different directions and can give rise to different grammatical markers depending on the syntactic surrounding.According to Lehmann (1992, p. 406, quoted in Traugott 2008, 221)  (IPI PAN) On the other hand, the same verb gave rise to a different grammatical marker when used with a prepositional phrase containing a deverbal noun as in: We see that a single lexical verb served as the source for four different grammatical cxns.This points to the phenomenon of polygrammaticalization (cf.Lai 2001).An important aspect of the constructional view of gzn is the fact that meaning often cannot be attributed to single morphemes.It is, thus, evident that it is not the verb mieć which carries the modal, hypothetical or existential meaning, but the cxn as a whole.In section 4 we are going to illustrate this constructional approach on the basis of Polish data.

Contact induced gzn
A further aspect that has recently spawned a series of studies concerns language contact as an important factor triggering gzn processes.A seminal work is Heine and Kuteva (2003), later expanded in Heine and Kuteva (2005), where the authors elaborate a theory of contact induced grammaticalization.They distinguish between two types of contact-induced gzn: 'replica contact induced gzn' and 'ordinary contact induced gzn'.As the authors present only scarce data from Polish we will try to put forward some hypotheses about how this model could be applied to the question of the influence of German influence on Polish.As there is a striking structural convergence, we can assume that the gzn of the do+deverbal noun was triggered by the German model (see Weiss 1987Weiss , 1988)) Heine and Kuteva (2005, p. 81) we can call this change a 'replica contact induced grammaticalization' involving the following steps: 1. Speakers of Polish notice that speakers of German use the possessive verb and the copula with a deverbal noun introduced by the preposition with a directional meaning.2. They develop an equivalent cxn using material available in their own language.3. The speakers replicate the grammaticalization process they assume to have taken place in German, using an analogical formula; i.e. they use the directional preposition do and the deverbal noun in order to develop the new cxn.
This type of contact induced gzn has to be distinguished from what Heine and Kuteva call 'ordinary contact induced gzn'.They differ in the framing of part 3: in the latter case, the speakers do not try to copy the gzn process they assume to have taken place in the source language, but draw on universal strategies of gzn (Heine and Kuteva 2005, p. 92).We would like to claim that the development of the distinction between strong and weak necessity in the field of modality as expressed by the dedicated Polish modals mieć/powinien vs musieć could be analysed in terms of 'ordinary contact induced gzn'.As Hansen (2005) in his areal linguistic study on the distribution of modal cxns among the languages of Eastern Europe shows there is a cline in the spread of the distinction expressed by the German modals müssen 'must/have to' vs sollen 'should, to be supposed to'.Both are polyfunctional modals (see Section 4.1) which differ in the expression of the strength of the necessity or obligation.As müssen expresses a strong and sollen a weakened necessity, the negation of the former is compatible with the positive form of the latter; cf.: (13) Diese Kartierung soll, muss aber nicht thematisch mit der Diplomarbeit verknüpft sein.‚This field mapping should, but need not necessarily be linked to the diploma dissertation.'(http://geostudium.uni-goettingen.de/) On the basis of the situation in Old Church Slavonic as described in Pallasová (1991), Večerka (1996) and Hansen (2001) we can assume the distinction between strong vs weak necessity to be the result of a later development in the individual Slavonic languages.According to Hansen (2000Hansen ( , 2005) ) this distinction which is quite typical of Germanic is attested in dedicated modals in Upper Sorbian, Czech, Slovak, Polish, Slovenian, Polish/Croatian, but not in Bulgarian or Russian. 7The following example from modern Polish shows that mieć (also powinien) and musieć can be used in a similar way as their German counterparts: (14) Ocena stopnia opanowania sprawności czytania ze zrozumieniem ma (ale nie musi mieć) zazwyczaj formę testu pisemnego i daje się łatwo wymierzyć punktami.(www.euro-edukacja.pl)(15) List motywacyjny w odpowiedzi na ogłoszenie powinien (ale nie musi) w nagłówku zawierać informacje o naszej osobie [...] (http://cv.wzory.biz/) Rewording the passage from above we can assume the following scenario of an ordinary contact induced gzn: 1. Speakers of Polish notice that speakers of German distinguish strong and weak necessity by using dedicated modals for each notion.2. They develop equivalent cxns using material available in their own language.3. The speakers draw on universal strategies of grammaticalization by mapping this distinction onto the modals mieć vs musieć which is not an analogical formula because sollen etymologically goes back to the meaning 'to owe' and mieć to 'to have'.
4 Gzn phenomena in Polish: two case studies

The gzn of modal cxns
In Polish grammaticography modals are usually treated as lexical verbs.In Germanic, but also Czech, Slovak and Serbian linguistics, however, there is a tradition to treat modal verbs as a category distinct from lexical verbs. 8Correspondingly, Czech and Slovak grammarians use the term vlastní modální (způsobová) slovesa / modálne čiže spôsobové slovesá 'proper modal (auxiliary) verbs / modal or auxiliary verbs' and describe the expressions of necessity, possibility and volition as part of the grammar (see for Czech Karlík et al. 1995, for Slovak Mistrík 1983, for Serbian Stanojčić/Popović 1998).This is perfectly in line with the view that modals are elements which have undergone a gzn process.They express the basic notions of 'necessity' and 'possibility' 9 and show syntactic properties of auxiliaries.From a typological point of view modals can be grasped in the following way: 7 These languages preferably resort to the conditional form of a modal of necessity as in Russian: dolžen (must.m.sg) vs dolžen byl by (must.m.sg.be.past.m.sg cond) 8 Studies on Modality in Slavonic based on gzn theory are: Hansen (1998 ff), Besters-Dilger et al. (2009), Ivanová (2009) and Roeder and Hansen (2006). 9The question whether modality comprises volition or not is still open.
A fully-fledged modal is a polyfunctional, syntactically autonomous expression of modality which shows a certain degree of grammaticalisation.'Polyfunctional' is understood as covering a domain within the semantic space of modality.A fully-fledged modal functions as an operator on the predicational and/or the propositional level of the clause.(Hansen and de Haan 2009, p. 512) Typical modals are polyfunctional in the sense that they express no less than two types of modality.One usually distinguishes dynamic, deontic and epistemic modality.Modals are polyfunctional, while so called modal content words, i.e. words with modal meaning which are not subject to a gzn process, have only one modal meaning.This difference could be illustrated on the basis of the word pair móc 'can' vs umieć 'to be capable'.The former can express 'capability', 'objective possibility', 'permission' (deontic), and 'medium degree of probability', while the latter is confined to 'capability'.
capability (dynamic): This semantic polyfunctionality can be seen as the result of an increased process of semantic bleaching.Research on Old Polish has shown that this polyfunctionality of modals grew in historical times; e.g. it could be shown that the epistemic meanings developed relatively recently (Hansen 1998 and2001): the first instances of the epistemic use of musieć are found in the second half of the 16 th century and epistemic powinien is not attested before the 19 th century: On the surface modals look like lexical verbs but syntactically they share properties with grammatical markers.The modal takes over the first semantic valence slot of the main verb and does not influence the selection of the subject argument.The following sentences show that the syntax of modals is different from non-grammaticalized content words; compare again móc with umieć: The modal móc does not effect the selection of the first argument and therefore allows impersonal sentences as in ( 23) or passive cxns (25).S tudent in (24) fills a semantic argument position opened by the verb rozwiązać 'to solve' and not by móc.The non-grammaticalized lexeme umieć 'to be capable' determines the selection of the first argument and therefore cannot be used in impersonal or passive sentences.This syntactic behaviour indicates that the modals open two syntactic (subject phrase and infinitive phrase), but only one semantic valence slots.The loss of a syntactic valence slot is a typical gzn symptom of the condensation of the structural scope.The gzn profile of modal cxns can, thus be summarized in the following way: modals arise from lexical verbs and become polyfunctional modal markers (semantic bleaching), in some cases they retain old lexical meanings (e.g. the modal wypadać), but do not underly phonological erosion.They show the loss of valence slots which is a symptom of a condensation process leading to an analytical verbal form.This specific gzn profile allows comparisons with other elements underlying gzn processes in Polish, but also in any other language.

The gzn of imperative cxns
In this paragraph, we would like to give a more comprehensive account of the gzn of the imperative marker niech (for details see Hansen 2004b).We are going to look into the gzn parameters erosion of form, coalescence with the lexical verb and fixation of word order.In its typical usage niech expresses a request directed towards the third or -in formal address -the second person: (26) Niech Piotr spróbuje tego pasztetu!(26') Niech pan spróbuje tego pasztetu!According to Havránek (1980, p. 462ff) and Brückner (1927, p. 360) the uninflected morpheme niech etymologically goes back to the negation of the negated imperative of the verb hajati 'to bother'. 10The Polish verb niechać still attested in Old Polish (see ex. 2 above) got out of use and has been replaced by the prefixed verb zaniechać.Niechaj got shortened to niech.So we find a sequence of shortening processes: nie chaj > niechaj > niech.Now, we will briefly look into the gzn parameters 'bondedness' and 'fixation'.The data show that the imperative marker, indeed, tends to be located in juxtaposition with the lexical verb.It is, however, still possible to put lexical items between imperative marker and main verb; these are e.g.clitical pronouns or adverbs as in the following examples: (27) Niech pan tak nie mówi!A second question is whether the imperative markers occupy a fixed position in relation to the main verb.The most obvious restriction concerns the position before or after the main verb.Although there are no strict rules concerning juxtaposition we find a strong rule determining the word order 'imperative marker precedes main verb'.The imperative marker tends to be placed at the beginning of the clause, but it can also occupy the second syntactic slot after the the first constituent or a particle (so-called Wackernagel position); cf.: To sum up, in the case of niech we find clear symptoms of erosion of the phonological substance, shifts from more concrete to more abstract meanings, a medium degree of coalescence with the lexical verb and a strong fixation of the sequence imperative marker before lexical verb.

The gzn of avertive cnx
Without going into details we would like to mention a gzn path which recently attracted the attention of the general linguists.As in many other languages (e.g.Russian.Serbian/Croatian, German) the verb chcieć gave rise to the usage labeled as 'opierać się czemuś, przychodzić z trudnością' (Słownik języka polskiego 1992).In the typological literature, this function is called 'avertive' (Kuteva 1998;see Drobnjaković 2009).The avertive gram indicates that an action was on the edge of taking place but eventually did not: (30) Mam subaru libero i jest taki problem, że nie chce zapalić.'I've got a Subaru and I've the problem that it doesn't want to start.' (http://motoforum.pl) The avertive differs from the volitional meaning in its restriction to negative contexts and the compatibility with inanimate subjects.
5 Outlook: further perspectives on the study of gzn phenomena in Polish In the previous section we have presented analyses of selected Polish cxns in the light of gzn theory whereby we had to restrict ourselves to some very few aspects.As a matter of fact, gzn phenomena are ubiquitous, and, therefore we would like to conclude this article by listing some further areas of future research on gzn phenomena in Polish 11 : the development of secondary prepositions from other parts of speech like podczas, zamiast, według etc. -the transition of the perfect tense into a general past tense as in the change przysiągł jeśm > przysiągł-em. 12-The rise of new imperative markers derived from wziąć in spoken language as in (31) Weźcie się ze mnie nie śmiejcie.
- the rise of the resultative cxn with mieć + particple as in mam to zrobione 14  the rise of the passive cxn with zostać This non-exhaustive list shows the perspectives of a fruitful application of gzn theory to the panchronic study of the Polish language.
'[G]rammaticalization does not merely seize a word or morpheme[. . .]  but the whole construction formed by the syntagmatic relations of the elements in question'.One case in point would be the different gzn paths of the Polish possession verb mieć which originally formed a transitive cxn with the possessor encoded in a nominal phrase in the nominative case and the possessee expressed by a noun phrase in an oblique case: Possessive cxn: [NP] N om [ma [NP] Acc ] (6) Przecież to nie on dzwoni, on ma klucze (IPI PAN) 6 On the one hand, mieć showed up in new syntactic contexts in which it developed into a marker of modality (strong necessity) and further into a marker of hypothetical assumption (for details see Hansen 2009); cf.: Modal cxn [NP] N om [ma [InfP]] Necessity marker: (7) Wokulski zawahał się, co ma zrobić w tej chwili.(IPI PAN) Mood marker: (8) Cele zjazdu są zbyt ważne, żeby coś takiego miało zakłócić jego przebieg.