anna młynarczyk-aalstein - aspectual pairing

7/29/2019 Anna Młynarczyk-Aalstein - Aspectual Pairing

http://slidepdf.com/reader/full/anna-mlynarczyk-aalstein-aspectual-pairing 1/3

Aspectual Pairing in Polish

Anna Aalstein and Patrick Blackburn

The received view of Slavic aspect is that it is intrinsically complex, and thatthere is little hope of discerning any substantial regularity. We argue that asfar as Polish aspect is concerned, this view is mistaken. We argue that the vastmajority of Polish verbs really do occur in aspectual pairs and that far frombeing a mysterious process, aspectual pairing in Polish is simple and regular.

We present an aspectual classification of Polish verbs that pins down themechanisms underlying aspectual pairing in Polish. Our classification startswith the observation what is usually (pre-theoretically) regarded as a Polishaspectual pair consists of two verbs, one of which is basic, with the other onederived from it by means of a ‘formant’. Formants are always one of the fourtypes of affix given below. Our classification is formationally-driven: we dividePolish basic verbs into five classes on the basis of the patterns of aspectualaffixation they enter into. That is, we classify Polish verbs according to whetherthey can undergo empty prefixation, delimitative poprefixation, semelfactive-na- suffixation, or morphonological change.

What motivates this particular choice of affixes as formants? In the litera-ture, there is much discussion of what a legitimate aspectual pair is (and what

merely qualifies as Aktionsartal variants). Our criterion is formal: every derivedverb that passes our Secondary Imperfectivisation test (and its mirror image,the Secondary Perfectivisation test) qualifies as an aspectual counterpart of agiven basic verb, and the affix involved qualifies as a formant. Our SI test is amodified version of the well-known Slavic SI test. Briefly, the idea underlyingthe SI test is that if you cannot undo the perfectivisation process in any otherway than by going back to the original basic verb, then you have formed a gen-uine aspectual pair, and the perfectivising affix you used is a formant. WhenPolish verbs are so classified, they group into five basic classes, as illustrated bythis table:

ep po- -na- mpc

class1 yes

class2 yesclass3 yes yesclass4 yes yes yesclass5 yes

(In the top line of the table, ep stands for empty prefix, po- is the delimitativeprefix po-, -na- is the semelfactive suffix -na-, and mpc stands for morphono-

1



logical change, that is, vowel change and/or suffixation of a perfective verb, for

example insertion of the -yw- suffix. For readers unfamiliar with Polish, weremark that all verbs in Class1 through Class4 are imperfective, and that allverbs in Class5 are perfective.)

In spite of its essentially formal nature, the classification reveals consider-able semantic regularity in the Polish verb system. Indeed, the classificationinduces temporal distinctions on Polish verbs, distinctions that look rather likeVendler-style distinctions. In particular, we claim that each of five basic classesis a semantically coherent group of verbs: class1 verbs are either state verbs orgradual transition verbs; class2 verbs are process verbs; class3 verbs are culmi-nating process verbs; class4 verbs are unitisable process verbs; and class5 verbsare culmination verbs.

We can formally model the temporal semantics of the verbs in each of theseclasses within the framework of event semantics. Space limitations rule out a

discussion of all five classes, so we confine ourselves here to some brief remarkson the verbs in Class5 and Class3, which we hope will convey the flavour of our approach. In what follows, we assume the reader is familiar with λ-drivensemantic construction. We use @ to denote functional application.

Here are the semantic entries for a typical verb in Class5 (we have chosenkupic) and it’s aspectual twin kupowac:

kupic: to buy, perfective formλwλy(w@λzλsλe(kupic(e) ∧ agent(e, y) ∧ patient(e, z ) ∧ conclusion(e, s)))

kupowac: to buy, imperfective form: mpc(kupic)

λwλy(w@λzλsλe(kupic(e) ∧ agent(e, y) ∧ patient(e, z ) ∧ induration(e, s)))

(The annotation mpc(kupic) indicates that kupowac is formed from kupic

by morphonological change.)

Note that the representation of the paired verbs kupic and kupowac isidentical, save that for the perfective form kupic the evaluation time s (whichwe model as a point of time) is asserted to be the time at which event e finishes(the conjunct conclusion(e, s) forces this), whereas in its imperfective twinkupowac the evaluation time is asserted to belong to the run-time of event e

(the conjunct induration(e, s) forces this).Let’s now look at a typical verb in Class3. Here matters are more interesting.

According to our classification, each verb in Class3 gives rise to two distinctaspectual twins, one formed by empty prefixation, the other by delimitative poprefixation. Let’s look at an example:

pisac: to write, imperfective formλwλy(w@λzλsλe(pisac(e) ∧ agent(e, y) ∧ patient(e, z ) ∧ induration(e, s)))

napisac: to write, perfective form: ep(pisac)λwλy(w@λzλsλe(pisac(e)∧agent(e, y)∧patient(e, z )∧conclusion(e, s)∧culm(e)))

popisac: to write, perfective form: po(pisac)

λwλy(w@λzλsλe(pisac(e)∧agent(e, y)∧patient(e, z )∧conclusion(e, s)∧¬culm(e)))

(The annotation ep(pisac) indicates that napisac is formed from pisac by

2



empty prefixation. The annotation po(pisac) indicates that popisac is formed

from pisac by delimitative poprefixation.)As with our previous Class5 example, the imperfective form (here pisac)is semantically distinguished by the use of the induration predicate from theperfective (note that its perfective twins napisac and popisac both have theconclusion predicate in their representations). But the semantics of the twoperfective twins is then further subdivided by a finer-grained specification onthe nature of e: the distinction between culminating processes that reach theirculmination, and culminating processes that are broken off. This distinction(here indicated by culm(e) and ¬culm(e) respectively) is what underlies thedifference of meaning between perfective verbs formed by empty prefixation of Class3 verbs, and those formed by delimitative poprefixation. This distinctionhas interesting repercussions for the imperfective paradox, but space limitationsrule out further discussion of this point here.

In the full version of this paper we shall motivate the methodology underly-ing our verb classification (in essence, it rests on systematic use of the SecondaryImperfectivisation and Secondary Perfectivisation tests), give a full account of the semantics of all five verb classes, and demonstrate that our system is compo-sitional. We shall also relate our work to traditional accounts of Polish aspect,and in particular to Czochralski’s classic study of the Polish aspectual system.1

Roughly speaking, our classification based approach treats as genuine aspectualtwins certain verb pairs that Czochralski views as Aktionsartal variants. Ourclassification sheds a new light on the much discussed distinction between Slavicaspect and Aktionsart. Crucially, it reveals that this distinction is not a cat-egorial one, but stems from the aspectual class of the particular (basic) verbthat is perfectivised/imperfectivised. For instance, our classification shows that

what Czochralski (following Isacenko and Maslov) calls ‘resultative Aktionart’,can only arise by making a Class3 verb perfective by means of an empty pre-fix, and that what he calls ‘inchoative Aktionsart’ arises when making a Class1verb perfective. On the other hand, what he argues is a true aspectual pair,is simply a basic Class5 verb together with its counterpart derived by meansof mpc. To put it another way, different classes of verbs simply have differentaffixational possibilities. There is no way of (say) applying an empty prefix to aClass5 verb, or of applying mpc to a Class3 verb. It follows (in contrast to howthis issue is commonly thought about) that an empty prefix is not a (universal)means of deriving a resultative Aktionsart, nor is mpc the sole mechanism forforming ‘true’ aspectual pairs. Rather, the vast majority of Polish verbs occurin aspectual pairs, and the form these pairs take depends on the class to whichthe basic verb belongs.

1Jan Czochralski, Verbalaspekt und Tempussystem im Deutschen und Polnischen. Eine

konfrontative Darstellung. Warszawa: Panstwowe Wydawnictwo Naukowe, 1975.

3

anna młynarczyk-aalstein - aspectual pairing

Documents