The proposed paper examines the clash between two central features of the Old Russian (OR) grammatica system: pro-drop and 2nd position clitics. This clash has consequences for the analysis of Slavic clitic systems that have not been previously observed in the literature. Primary material is drawn from OR texts (our data and general analytical framework are taken from Choo 2003), and then we compare OR to parallel in modern Slavic languages with similar structural features (general theoretical background based on Franks and King 2000).
OR through the 16th century was a canonical pro-drop language (like the contemporary West and South Slavic languages). Therefore the grammatical subject, when pronominal, is regularly null unless emphasis, contrast, or other pragmatic factors force it to be overt.
(1) a. I __ pride v" slověni, ... 'And he reached the Slavs...'
b. #I on" pride v" slověni, ...
[Primary Chronicle, 26]
Thus, the attested example (1a) is normal, while (1b) is anomolous unless there is a discourse reason to overrule the structural preference for pro-drop (# means anomolous in the supposed context).
At the same time, OR has a rich system of clitics (pronouns, auxiliaries, li, že, bo, etc.) which stand in 2nd position in the sentence, following the first stressed word or phrase.
(2) a. Igor" že nača knjažiti v" Kyevě...
‘Thus Igor' began to rule in Kiev...’
b. *__ že nača knjažiti v" Kyevě...
c. *Igor" nača knjažiti že v" Kyevě... (Primary Chronicle, 68)
Thus, in the attested (2a) the clitic cluster follows the initial stressed NP ....., whereas variants such as (2b) or (2c) never occur in our OR corpus.
We must also note that OR respects the general functional organization of the sentence familiar from the modern Slavic languages (we assume a generic theory of functional sentence perspective (FSP) as in Ковтунова 1976). Thus, thematic/old/given/background/generic elements come at the beginning of the sentence, whereas rhematic/new/communicatively dynamic elements come toward the end.
So the question arises: shouldn't these two structural principles come
into conflict in some possible sentences? What happens if a short sentence contains
a null pronominal subject and one or more clitics and there
is no other element in the sentence such is
communicatively suited for thematic fronting (such as an adverb or prepositional
phrase)? In this abstract situation something has to give: a) the pronominal
subject could become overt (i.e., neutral pro-drop is violated in order to prevent
a violation of clitic-second); b) the clitics could stand in initial position
(violating clitic-second in order to preserve pro-drop); or c) some other element
could be fronted (disrupting the standard FSP organization of the sentence but
preserving pro-drop and clitic-second).
In OR there are attested examples which can be argued to instantiate strategy (a), as in (3):
(3) I __ ochjuti plach', i reče: "Kdě se __ esm"?" Oni že rekoša emu: "V" Zviždeni gorodě." I __ vprosi vody, oni že daša emu, i __ ispi vody, ...
‘He [Vasil'ko] heard [her] weeping, and said “Where am I?” THEY replied "Zvižden." And he begged for water, and THEY gave him some, and he drank some water, ...’ (Primary Chronicle, 252)
But when informants for various modern Slavic languages are presented with contexts like (4), where this conflict arises starkly, they inevitably opt for strategy (c), as in (5):
(4) [Context] What did John do to his sister?
(5) a. Obidi ja. ‘[He] offended her.’ (Bulgarian)
b. #Toj ja obidi.
c. *Ja obidi.
The preferred strategy in (5a) involves fronting the verb, which is the rhematic focus of the answer. The clitic then follows it, even though it is no part of the rheme of the sentence. The specific import of this example depends upon the way one analyzes clitics (functional or lexical elements), but at the very least it does offer a surface disruption of neutral FSP sequencing.
A formal analysis of strategy (c) will be offered in terms of Optimality Theory, in which three constraints have the following relative rankings:
(6) Clitic-Second > Pro-Drop > FSP
We argue that this ranking reflects an essentially psycholinguistic fact of Slavic languages, based on the degree of subjectivity in the three phenomena:
Clitic-Second: a rigid, inviolable ordering constraint; no speaker subjectivity involved; no competing alternatives;
Pro-drop: subjectivity involved in terms of degree of emphasis, but it is a rather straightforward threshold choice based on a relatively simple, intuitive factor (witness the striking uniformity in pro-drop usage across the Slavic languages); only two alternatives (null/overt) available to speakers;
FSP: Highly subjective; subject to influence from multiple factors involving discourse, real-world circumstances, speaker psychology, etc.; multiple alternative word-orders.
We then return to other examples from OR which instantiate strategy (c), but with a richer supply of predicate elements, and show that OR patterns exactly like the modern Slavic languages in respecting the ranking in (6).
References
Choo, Sukhoon. The Decline of Null Pronominal Subjects in Old Russian. Ph.D. dissertation, Indiana University, 2003.
Franks, Steven and Tracy Holloway King. A Handbook of Slavic Clitics. New York: Oxford University Press, 2000.
Ковтунова, И. И. Современный русский язык: Порядок слов и актуальное членение предложения. Москва: Просвещение, 1976.