At the 2008 annual conference of the SLE in the Italian provincial town of Forlì, Andrej Malchukov and Anna Siewierska organized a workshop on “Impersonal constructions: a cross-linguistic perspective”. In August 2011, a 641-page volume was published by Benjamins, edited by the two workshop organizers. (Sadly, the publication of the book coincided with Anna Siewierska’s tragic death on 6 August 2011.) This is the first book on impersonal constructions cross-linguistically, and its 20 papers cover a wide range of languages (over 150 languages and families are listed in the index).
I have long wondered what the factor uniting the various phenomena traditionally called “impersonal” might be. The book has little discussion of the terminological tradition, and it does not seem to ask who first proposed that they should be treated together. According to the OED, the term goes back at least to Palsgrave (1530), and it was first used for verbs that do not vary for person (and number), but occur only in the 3rd person singular, such as Latin pluit ‘it is raining’. (Nowadays we would say that these are verbs that have no subject, or only an expletive subject.) If the restriction to 3rd person singular forms was the original intuition, then modal verbs such as Latin oportet ‘it is necessary’ would also have fallen in this class. Later we find collocations such as “impersonal passive” (referring to Latin passives such as Latin agitur ‘one acts, lit. it is acted’, or German es wird getanzt ‘one dances, lit. it is danced’) or “impersonal pronoun” (referring to French il as in il vient trois personnes ‘three people have come’).
The present book covers a very hetergoneous range of phenomena which are loosely bound together by the fact that they are similar to pluit, oportet, agitur or il vient in one way or another:
- (A) meteorological constructions (e.g. in Uralic languages, M. Salo; in Oceanic languages, C. Moyse-Faurie)
- (B) passive-like constructions in which no direct object is promoted to subject position, so that they are subjectless (“impersonal passives”, e.g. in Ainu, A. Bugaeva)
- (C) presentational constructions like Spanish hay dinero ‘there is money’ (e.g. V. Gast & F. Haas)
- (D) “generalized-man” constructions, like on in French and man in German (e.g. Siewierska on European languages, and Yan & Siewierska on Mandarin)
- (E) “generalized 3rd-person plural” constructions, like In Spain, they eat late (e.g. Siewierska)
- (F) “generalized 2nd-person” constructions like You wouldn’t get away with such a remark here.
- (G) “generalized subject-reference” constructions with a special form, such as si in Italian (In Italia si mangia spaghetti ‘In Italy one eats spaghetti’).
- (H) zero-subject constructions with a “generalized” interpretation (e.g. Mandarin Ø yīnggāi jìnkuài líkāi [Ø should immediately leave] ‘One should leave immediately’, Yan & Siewierska)
- (I) experiential constructions with no nominative subject, as in German mich friert [me freezes] ‘I am freezing’ (e.g. Verstraete on Australian languages)
- (J) one-place predicates that combine with embedded clauses as subject, e.g. It is good that you came, Playing chess is fun.
One can see that all of these share some properties with pluit/oportet/agitur/il vient, but is that sufficient to call them all “impersonal constructions”? Using a technical term normally implies that the term has a clear meaning, but what exactly an impersonal construction is supposed to be does not become clear.
Originally, the term referred to a formally defined category: Impersonal verbs are verbs that occur only in the third person (e.g. Götzinger 1836: 308) or verbs that lack a subject. But for quite some time it has also been common to refer to constructions like (D)-(F) as “impersonal”, although these share no formal properties with the others. This is, incidentally, a frequent type of semantic change in linguistic terminology: A term for a formally defined category takes on a semantic sense when it is carried over to another language that lacks the formal properties of the first language (as nicely shown by Lehmann (2007), a brilliant but almost unknown article on semantic change in linguistic terminology).
Once “impersonal” takes on a semantic flavour (and is seen as whatever impersonal passives like es wird getanzt and general pronouns like one in one dances share), one can even talk about “impersonality”, and ask how “impersonality” is expressed, i.e. how (some of) the senses of A-J are rendered in a given language. But no definition of this semantic notion that encompasses all of A-J, and only these, seems to be possible. Thus, “impersonal constructions” are both formally and semantically so heterogeneous that there is no way to say what is an impersonal construction and what isn’t (or even what a “canonical” impersonal construction is and what a less canonical one is).
In the introduction to the volume, Malchukov & Siewierska recognize the heterogeneity of the constructions, but suggest that what unites them is the “lack of a (referential) subject”. If the “generalized-subject” constructions in (D)-(G) are regarded as having a non-referential subject, that includes all of (A)-(I), and for (J) (the Playing chess is fun type), one might also say the subject is not referential (hoping that “referentiality” is more clearly defined than impersonality). But what about generic-subject sentences like Lions are dangerous? Are these impersonal, too?
A tripartite subclassification that the editors suggest is into R-impersonals (with nonreferential subjects, i.e. meteo-verbs (A) and man-constructions (D)), T-impersonals (with non-toptical subjects, i.e. presentationals (C)), and A-impersonals (with non-agentive “subjects”, i.e. experiential constructions (I)), but not much is made of this. Instead, the major claim of the Malchukov & Ogawa paper is that the various uses of impersonals can be arranged in a semantic map on the basis of shared formal properties. This is of course an interesting result, which does not presuppose that the target domain be coherent either formally or functionally.
Thus, the editors seem to be happy that impersonality remains a vague “umbrella term” (to use a label that Zwicky 1994 introduced to characterize the term “clitic”). Be that as it may, the 20 papers in this volume are an interesting and important contribution to our knowledge of linguistic diversity in this area.
Götzinger, Max W. 1836. Die deutsche Sprache und ihre Literatur. Band 1. Die deutsche Sprache. Stuttgart: Hoffmann.
Lehmann, Christian. 2007. On the upgrading of grammatical concepts. Moerdijk, Fons & Santen, Ariane van & Tempelaars, Rob (eds.). Leven met woorden: Opstellen aangeboden aan Piet van Sterkenburg. Leiden: Brill, 409-422. (http://christianlehmann.eu/publ/upgrading.pdf)
Malchukov, Andrej & Siewierska, Anna. 2011. Impersonal constructions: a cross-linguistic perspective. Amsterdam: Benjamins, 641 pp.
Malchukov, Andrej & Ogawa, Akio. 2011. Towards a typology of impersonal constructions: a semantic map approach. In: Malchukov & Siewierska (eds.), 19-56.
Malchukov, Andrej & Siewierska, Anna. 2011. Introduction. Malchukov & Siewierska (eds.), 1-15.
Palsgrave, John. 1530. L’esclarcissement de la langue francoyse. London.
Zwicky, Arnold M. 1994. What is a clitic? In: Joel A. Nevis, Brian D. Joseph, Dieter Wanner, & Arnold M. Zwicky, Clitics: a comprehensive bibliography 1892-1991, xii–xx. Amsterdam: Benjamins.