Accent on accentless words

Before I came to concern myself with negative concord, I was at one point thinking a lot about accent shift to proclitics in Serbo-Croatian. I was reading a lot about the Neo-Štokavian(1) accent system in general (including a Serbian translation of Lehiste and Ivić, 1986, which was, unlike the original, available through my university’s Slavic Studies department), but I didn’t find a whole lot about the accent shift specifically back then, and least of all works. Two relevant works have appeared since: Werle (2009), a dissertation analysing the phenomenon from a mostly phonological perspective, and a short paper by Aljović and Riđanović (2009). I’m writing this post because last week, I finally got my hands on the latter.

I’ll try to make this as accessible as possible to people who haven’t previously concerned themselves with proclitics and/or Serbo-Croatian intonation, so bear with me while I attempt to explain a few terms.

A clitic is a word that doesn’t have an accent of its own and relies on “leaning” onto another, accented, word – its host – in pronounciation. A proclitic is a clitic that leans on the following rather than the preceding word – in Serbo-Croatian, these are mostly the negative particle “ne”, most prepositions, the conjunctions “i” (“and”) and “ni” (“neither”, “not even”). If that were all there is to say, we should expect (pro)clitics to be always unstressed. This is not the case in (certain varieties of) Serbo-Croatian, though.

Here, we need a brief excursus on the Neo-Štokavian accent system. The language is traditionally described as having for different accents: short falling, short rising, long falling, and long rising. Abstracting away from length, this yields two different accent qualities. However, this does not mean that any string of n syllables has 2n different possible accent patterns – indeed, for strings of n syllables, there are exactly n possible accent patterns: Falling accent on the initial syllable, and rising accent on any non-final syllable. Final accent on words with more than one syllable is excluded, as is falling accent on a non-initial syllable. This has given rise to an analysis of Neo-Štokavian accent in terms of “accent spread”, exemplified by Inkelas and Zec (1988). Here, the claim is that perceived accent in Serbo-Croatian does not directly reflect lexical accent, but rather that the correspondence is mediated by phonological rules. According to this analysis, any syllable (or none) can be lexically accented. Where lexical accent (realised as a high tone) is present, it spreads to the left if applicable; where no lexical accent is present, a default rule inserts accent on the leftmost syllable. “Rising accent” is a syllable associated with high tone, followed by another high-toned syllable, and “falling accent” is a syllable associated with high tone, followed by a toneless (and thus by default low) syllable. This gives the following possibilities:

  1.  lexical high tone on a non-initial syllable n: tone spreads to syllable n-1, n-1 and n both carry high tone, n-1 is perceived as carrying a “rising accent”.
  2.  lexical high tone on the initial syllable: no spread is possible, the initial syllable is the only one carrying tone and perceived as carrying a “falling accent”.
  3. no lexical tone: default rule inserts tone on initial syllable, proceed as in (2).

It follows from this and similar analyses that a proclitic will receive stress in conditions (2) and (3) if and only if the collocation proclitic+host is interpreted as one word by the phonological component. Indeed, that’s what we sometimes find, most frequently in the “classical” or “Vukovian” perceived standard, as well as in regions of Bosnia-Herzegovina and Montenegro, much less in the modern colloquial of Belgrade or Zagreb where accent shift (other than to “ne”) is essentially limited to fixed idioms. Other than geography and idiomatic status, factors influencing the distribution of accent shift include length of the host and “[strength of] semantic or syntactic connection”(2). My approach would have been to attempt to formally characterise the latter condition: Focusing on a dialect that retains productive accent shift and keeping frequency and host length constant, what are the properties that unify contexts in which the shift fails to occur, i.e. is this the consequence of additional syntactic structure intervening – Aljović and Riđanović (2009) list examples where the proclitic is followed by coordinate structures, post-modified nouns, and appositive pre-modifiers as contexts where accent shift essentially never happens?

The data are quite messy, with all those interfering phonological, quasi-phonological (host length), and frequency-based effects, so I guess I can call myself lucky to have found something else to write about. For now.


(1) Neo-Štokavian is the term preferred by dialectologists to refer to a group of South Slavic dialects that includes (the “classical” variants of) the modern standard languages Bosnian, Croatian, Serbian, (and Montenegrin), and that display the shifted accent pattern. Serbo-Croatian is used almost interchangeably as a cover term for those languages.

(2) “[Transfer failure] can result from a weaker semantic or syntactic connection, but also from the lesser frequency of a syntagm, or from the greater length of its parts. [Transfer] is most consistently realized with verbs in connection with ne… it is very common for pronouns with prepositions (and with the conjunctions i and ni), with nouns it is almost completely limited to prepositional syntagms… and is least extensive with adjectives” – Ivić (1965), quoted and translated in Werle (2009).

Aljović, Nadira and Midhat Riđanović, 2009: On the Shift of Bosnian Accents from Host to Proclitic: New Insights. Steven Franks, Brian D. Joseph, Vrinda Chidambaram, Wayles Browne (Eds.): A linguist’s linguist: studies in South Slavic linguistics in honor of E. Wayles Browne. Indiana: Slavica.

Inkelas, Sharon and Draga Zec. 1988. Serbo-Croatian pitch accent: The interaction of tone, stress, and intonation. Language 64:2. 227-248.

Ivić, Pavle. 1965. Prozodijski sistem savremenog srpskohrvatskog standardnog jezika. Stanisław Drewniak, ed. Symbolae linguisticae in honorem Georgii Kuryłowicz. Prace Komisji Językoznawstwa, Nr 5. Wrocław: Polska Akademia Nauk. 135-144.

Lehiste, Ilse and Pavle Ivić. 1986. Word and Sentence Prosody in Serbocroatian. Cambridge: The MIT Press.

Werle, Adam, 2009. Word, Phrase, and Clitic Prosody in Bosnian, Serbian, and Croatian. PhD Dissertation, University of Massachusetts.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s