Slavic languages and (generative) phonology

Setting the scene

What counts as phonology?

• Distributions
• Allophonic alternations
• Neutralizing alternations
• Morphophonology

Distributions and allophony: Bulgarian vowels

Phoneme	Stressed	Unstressed	Alternation
/a/	ˈɡlaven ‘main’	ɡlɐˈva ‘head’	[a] ~ [ɐ]
/ɛ/	ˈvrɛme ‘time.SG’	vremeˈna ‘time.PL’	[ɛ] ~ [e]
/ɔ/	ˈkɔsəm ‘hair’	kosˈmat ‘hairy’	[ɔ] ~ [o]
/i/	ˈtip ‘type’	tiˈpɤt ‘type.DEF’	[i]
/u/	ˈkupʲə ‘buy.PRS.1SG’	kuˈpuvɐm ‘buy.IPFV.PRS.1SG’	[u]
/ə/	ˈɫɤk ‘bow’	ləˈkɤt ‘bow.DEF’	[ɤ] ~ [ə]

For the traditional view, see Tilkov (1982); for a detailed recent empirical study, see Sabev (2023).

The pairs [a] ~ [ɐ], [ɛ] ~[e] etc. are in complementary distribution determined by stress. In structuralist phonology, we say they belong to the same phoneme; the basic variant is the one found in stressed syllables.

In generative phonology, we say that the unstressed variants derive by a rule that applies in unstressed position. In stressed syllables, no rule applies and the vowel in the underlying representation surfaces unchanged. Because stress in Bulgarian is mobile, this machinery creates alternations where the same morpheme occurs in different shapes depending on where stress falls: [ɡlav] ~ [ɡlɐv], [mɔr] ~ [mor] etc. These are allophonic¹ alternations, because the alternating segments do not contrast underlyingly.

1. Or ‘non-neutralizing’

Neutralizing alternations: Bulgarian soft consonants

Most Bulgarian consonants come in hard (plain) and soft (palatalized) versions, which contrast before back and central vowels.

Following vowel	Hard	Soft
[a]	baɫ ‘ball’	bʲaɫ ‘white’
[ɤ]	ɡlɐˈsɤt ‘voice.DEF’	ɡlɐˈsʲɤt ‘read.PRS.3PL’
[u]	kup ‘pile’	kʲup ‘cauldron’
[ɔ]	poˈzɔr ‘shame’	poˈzʲɔr ‘poseur’

Word-finally, only hard consonants are allowed. Many lexical items alternate accordingly.

Word-final	Prevocalic	Gloss
sɤn	səˈnʲɤt	‘dream’
kɔn	ˈkɔnʲət	‘horse’
kraɫ	ˈkralʲət	‘king’
vaɫ	ˈvaɫət	‘rampart’
zvɤn	zvəˈnɤt	‘peal’
trɔn	ˈtrɔnət	‘throne’

These alternations involve the pairs [n] ~ [nʲ], [ɫ] ~ [lʲ], which are otherwise distinct phonemes in the language. These are neutralizing alternations. In this case, the outcome of the neutralization is unambiguous: it is the hard consonant. Structuralist schools differ in their analysis of these, but the critical property is that they are obligatory — that is, they occur whenever the context is met. For this reason they can be called automatic.

Morphophonological alternations: Bulgarian palatalization

Before the front vowels [ɛ i], the hard-soft contrast is neutralized. For most consonants, the outcome is analysed as hard.¹

Contrast	Neutralization	Gloss
bʲaɫa	bɛli	‘white.FEM.SG ~ PL’
banʲa	bani	‘bath.SG ~ PL’
nʲama	nemi	‘mute FEM.SG ~ PL’
zɛmʲa	zɛmi	‘earth.SG ~ PL’

Velar consonants neutralize to soft in the same context, with alternations.

Hard	Soft	Gloss
dəˈɡa	dəˈɡʲi	‘arc’
ˈbitka	ˈbitkʲi	‘battle’
ˈɔrex	ˈɔrexʲi	‘walnut’

But other [i ɛ] suffixes do different things.

Hard	Palatalization I	Palatalization II
vɤɫk ‘wolf’	ˈvɤɫt͡si ‘PL’	ˈvɤɫt͡ʃi ‘ADJ’
ˈpɔdviɡət ‘feat.DEF’	ˈpɔdvizi ‘PL’	podˈviʒen ‘movable’
siroˈmax ‘pauper’	siroˈmasi ‘PL’	siromɐˈʃija ‘poverty’

In structuralist phonology, such alternations are generally considered morphophonological² and usually considered to be different in kind to allophony and obligatory neutralization.

1. We will revisit this

2. Sometimes also historic, or non-automatic to distinguish from automatic neutralization.

Slavic phonology and phonological theory

All of these types of alternations are pervasive across the Slavic languages. A lot of Western phonological theory was developed in Slavic-speaking countries, and had to deal with all of these issues. I would argue that in fact these properties of the Slavic languages were a key, and perhaps underappreciated, driver in the development of many flavours of phonology — including generative phonology.

Slavic phonology or Russian phonology?

As we will see, a lot of theory development has centred Russian. This undoubtedly reflects the dynamics of political and cultural power, and we should reflect on that. Russian is also a relatively conservative language with regard to sound change, and that is in some respects an advantage in the context of phonology, for reasons that will become clear soon. To what extent Russian-focused scholarship has occluded the patterns is a theme we will keep returning to.

The beginnings: The Kazan School

• Credited with developing the phoneme
• Strong focus on alternations
• Polish and Russian front and centre

Jan Baudoin de Courtenay

Mikołaj Kruszewski

The St Petersburg (Leningrad) School

Lev Shcherba is usually credited for developing lexical distinctiveness as the basis for defining the phoneme. He worked in Russia, and focused a lot on Russian; he also had a strong background in Slavic phonology, having written a thesis on a Sorbian dialect. His primary interest was in phonetics: alternations and neutralization were not as central to his theoretical approach.

• The phoneme as the minimal unit of lexical contrast
• Slavic background (wrote his thesis on a Sorbian dialect), focus on Russian

Lev Shcherba

Moscow to Prague: Nikolai Trubetzkoy

Phonology and the Prague Circle

Members of the Prague Linguistic Circle, including Trubetzkoy and (especially) Jakobson themselves, had wide-ranging interests beyond phonology and indeed linguistics, but here we focus on the phonology part. For a book-length treatment of the Prague Circle, see Toman (1995); I also recommend Sériot (2014) for the intellectual background.

Nikolai Trubetzkoy

Among theoretical phonologists, Nikolai Trubetzkoy is widely seen as the founding father of classic structuralist phonology, for his Principles of phonology (Trubetzkoy 1939). His interests were strongly focused on Slavic philology generally: a new historical grammar of the Slavic languages was his unfinished magnum opus (see also Trubetzkoy 1954). Trubetzkoy’ intellectual background is linked to Moscow, where he trained before leaving the country.

A notable feature of Trubetzkoy’s approach is the strict separation between allophonic alternation, automatic neutralization (implemented by positing underspecified archiphonemes), and morphophonology, considered to be a separate component of grammar. The theory of non-automatic alternations remained relatively underdeveloped (but see Trubetzkoy 1934), but it is clear that for him it is different in kind from automatic phonology.

• Fully developed structuralist phonological theory
• Long-standing focus on Russian and Slavic more broadly
• Strict separation of phonological neutralization and historic alternations

Moscow to Prague to Harvard: Roman Jakobson

Another product of Moscow, Jakobson was Trubetzkoy’s close collaborator in Prague, where he stayed until the start of the war. He made significant contributions to the analysis of Russian, Czech, and Serbo-Croatian phonology within the Praguian mould. After the war, he moved to the US and developed his theory in new directions.

• Close collaborator of Trubetzkoy
• Early influential work on Slavic within the Prague mould
• After the war: shift away from analysing relationships within individual systems

Roman Jakobson

In earlier work, Jakobson, like Trubetzkoy, was interested in the relationship of elements within an individual system. This can be seen in his highly influential Remarks on the phonological development of Russian compared to the other Slavic languages (Jakobson 1929), which offered an early example of Praguian phonological analysis and has been hugely influential in the synchronic and diachronic phonological analysis of the Slavic languages ever since.

Later, Jakobson became interested in information theory and the notion of ‘efficiency’ (see Dresher & Hall 2022). This led him towards developing more abstract analyses that were not limited to observing surface patterns of contrast and neutralization. A key moment was the publication of Russian conjugation.

The Russian conjugation moment

The traditional analysis of Slavic verbal inflection posits two stems (the infinitive stem and the present-tense stem) for each verb. Their distribution across the paradigm is seen as purely morphological. How the shapes of the two stems are related is also not predictable: it is a lexical property of the verb.

Russian verbal paradigm: plakat’ ‘to cry’

Infinitive stem			Present-tense stem
Infinitive		plaka-t’	Present tense	1SG	plač-u
Past tense	SG.M	plaka-l		2SG	plač-e-š
	SG.F	plaka-l-a		3SG	plač-e-t
	SG.N	plaka-l-o		1PL	plač-e-m
	PL	plaka-l-i		2PL	plač-e-te
Past participle		plaka-vš-		3PL	plač-ut
		(za)plaka-nn-	Imperative	2SG	plač
			Present participle		plač-ušč-

• In some verbs, the infinitive stem is formed by adding a vowel to the present-tense stem: plakat’ ~ plaču
• In other verbs, the infinitive stem is formed by truncation of the present-tense stem: dela-t’ ~ delaj-u ‘do’
• In yet others, both end in a consonant: nes-ti ~ nes-u ‘carry’

Jakobson’s big generalization

The general case:

• The ‘infinitive’ stem ends in a vowel and precedes consonant-initial suffixes
• The ‘present-tense’ stem ends in a consonant and precedes vowel-initial suffixes

All forms are derived from a single stem by applying a sequence of rewrite rules, notably deletion of a vowel before a vowel and deletion of a consonant before a consonant.

Verb	Infinitive stem	Present-tense stem
‘cry’	plaka-ti → plakat	plaka-u → plaču
‘do’	delaj-ti → delat’	delaj-u → delaju
‘carry’	nes-ti → nesti	nes-u → nesu

Riga to Harvard to MIT: Morris Halle

Morris Halle worked under Jakobson, and some of his earliest work is directly developing Jakobson’s new approaches, such as his Old Church Slavonic conjugation. The approach based on rewrite was front and centre in his Sound pattern of Russian, a work now not much read, but often seen as a precursor to the Sound pattern of English.

• Another émigré from Eastern Europe (in this case Latvia), Jakobson’s pupil
• Earliest work developing the ‘new’ Jakobsonian approach
• One half of Chomsky & Halle (1968)

Morris Halle

Slavic conjugation

With the basic structure of the verbal paradigm being very similar across the Slavic languages, the Jakobsonian generalization was widely applicable to other languages. In applying the method to Old Church Slavic, Halle (1951) discovered more rules that unified apparently disparate stems. Crucially, what such rules often did was reproduce the diachronic development.

Verb	Stem	Infinitive	PRS.1SG	AOR.3SG
‘cut’	rěza-	rěza-ti	rěža-ǫ	rěza-s → rěza
‘know’	znaj-	znaj-ti	znaj-ǫ	znaj-s → zna
‘lead’	ved-	ved-ti → vesti	ved-ǫ	ved-e-s → vede
‘cook’	pek-	pek-ti → pešti	pek-ǫ	pek-e-s → peče
‘curse’	klьn-	klьn-ti → klęti	klьn-ǫ	klьn-s → klę
‘blow’	dъm-	dъm-ti → dǫti	dъm-ǫ	dъm-s → dǫ

Expanding the scope of phonology

Arguably the more famous innovation by Halle was the analysis of voicing assimilation in Russian, originally presented in a talk entitled On the phonetic rules of Russian and reproduced in Halle (1959).

Russian shows final devoicing and voicing assimilation of obstruents across word boundaries. In most cases, the alternations we observe involve segments that are otherwise different phonemes: they are automatic neutralizing alternations. The affricates [t͡ʃʲ] and [t͡s] lack phonemic voiced counterparts, but still undergo assimilation: for these segments, the alternations are allophonic. However, they are patently the same process, so splitting them into two alternations — one neutralizing and one allophonic — misses the generalization.

Gloss	Word-final	Prevocalic	Assimilation context	Alternation
‘cat’	kɔt	kɐˈtˠi	kɔd bˠi	/t/ ~ /t/ ~ /d/
‘code’	kɔt	ˈkɔdˠi	kɔd bˠi	/t/ ~ /d/ ~ /d/
‘night’	nɔt͡ʃʲ	ˈnɔt͡ʃʲi	nɔd͡ʒʲ bˠi	/t͡ʃʲ/ ~ /t͡ʃʲ/ ~ /t͡ʃʲ/ [d͡ʒʲ]

It is commonly assumed that this was the killer argument against the ‘taxonomic phoneme’, which convinced phonologists (in North America at least) that derivational accounts aiming to capture generalizations without regard to surface contrast or phonemic status are superior to earlier approaches.

This is a little bit of a myth: see Anderson (2000) on the contemporary reception of Halle’s work, Kiparsky (2018) on the fact that the duplication identified by Halle was well known prior to his analysis, and Dresher & Hall (2020) on the wider repercussions. Still, Halle’s analysis of Russian remains central to the history of the field.

The big analytical innovation

• Premise: the grammar consists of rewrite rules
• Premise: rewrite rules handle historical alternations, live neutralizing alternations, and allophony
• Premise: rules responsible for allophony manipulate phonological units
• Conclusion: rewrite rules manipulate phonological units

The classic generative approach to Slavic phonology

For the rest of the week, we will look at some specific examples of how Slavic languages were analysed by generative phonologists from the early 1960s onwards.

• Large premium on generality and economy
• Wide empirical scope: aim for a comprehensive analysis of all kinds of patterns within the same system
• In practice: underlying representations that reproduce earlier historical stage, rules that recap the diachronic development

Diachrony and synchrony

In many ways, this is not really surprising: today’s irregular morphology is often yesterday’s regular sound change, so an economical account that captures all the generalization often is the diachronic one.

This was a feature, not a bug! The fact that underlying representations stuck closely to earlier stages was treated as a discovery: URs changed more slowly than surface representations.

The received analysis

From Slavic phonology to generative phonology

• Further work on Slavic itself (Halle 1963; Lightner 1963; 1965)
• Programmatic development of the approach (Halle 1962)
• Incorporation of links to new syntactic theories through notions like immediate constituency analysis and the cycle (Halle 1963; Lightner 1967)
• Chomsky & Halle (1968) as the crystallization of the generative approach to phonology

Was SPE a paradigm shift?

SPE is often treated as ushering in a Kuhnian scientific revolution — a paradigm shift that made what came before it largely irrelevant to the conduct of ‘normal science’. Is it true?

From a Slavic perspective, SPE most certainly did not cut short the development of the Trubetzkoy — Jakobson — pre-SPE Halle line. In the guise of a ‘morpho-phonological’ approach, it remained arguably the mainstream view of phonology in Slavonic studies departments in North America for a long time, and remains strong even now. For some reflections from the Slavic side, see Flier (1974), Shapiro (1974), or Sussex (1976). For views from the generative side, see Bethin (2006) and Kavitskaya (2017).

All week, I will argue that most non-specialist phonologists who have dealt with Slavic have really dealt with the Jakobson-Halle-Lightner version of it, and this is something we should problematize more.

Lightner vs. Modern Standard Russian

To give you a sense of this approach, here is the general picture of the Russian vowel system in Lightner (1965).

If this looks familiar to you from Slavic historical phonology… you’re not wrong.

Surface vowel	Underlying vowel	Comment
[i]	/ī/
	/ū/	Fronting in some contexts
[ɨ]	/ū/
[u]	/ou/
	/eu/	Source of surface [Cʲu]
	/oN/	When nasal is syllable-final
[e]	/ĕ/	When not → [o] |
	/ĭ/	By the Lower rule for yers, when not → [o] |
	/ē/	When not alternating with o
[o]	/ŏ/
	/ŭ/	By the Lower rule for yers
	/ĭ ĕ/	When subject to e → o |
[a]	/ō/
	/ē/	In cases like kričat’ ‘shout’ /krīk-ē-tī/
	/eN iN/	When nasal is syllable-final; source of surface [Cʲa]

Here are some underlying representations:

SR	UR	Gloss
žog	gĭg-l-ŭ	burn-PST-SG.M
žgla	gĭg-l-ō	burn-PST-SG.F
pisat’	pīs-ō-tī	write-INF
pišu	pīs-ō-ĕ-m	write-PRS.1SG
načat’	nō-kĭn-tī	begin.PFV-INF
načnët	nō-kĭn-ĕ–t	begin.PFV-PRS.1SG
načinat’	nō-kīn-ō-tī	begin.IPFV-INF

And here is a real derivation:

1. Pre-cyclic rule: /(plōt+ī+ī)+m/ → ((plōt+ī+u)+m)
2. First cycle: (plōt+ī+u) → (plōt+ju) → (plōtʲ+j+u)
3. Second cycle: (plōtʲ+j+u+m) → (plōtʲ+ū+m) → (plōtʲ+j+ū) → (plōčʲū) → [plačʲu]

Although this seems extreme, the basic idea proved remarkably resilient. Tomorrow, we look at consonant palatalization.

References

Anderson, Stephen R. 2000. Reflections on “On the phonetic rules of Russian.” Folia Linguistica 34(1–2). 11–28. doi:10.1515/flin.2000.34.1-2.11.

Bethin, Christina Y. 2006. Slavic phonology in the United States. (Ed.) Steven Franks, Edna Andrews, Ron Feldstein & George Fowler. Glossos 8. http://slaviccenters.duke.edu/uploads/media_items/8bethin.original.pdf (29 June, 2017).

Chomsky, Noam & Morris Halle. 1968. The sound pattern of English. New York: Harper; Row.

Dresher, B. Elan & Daniel Currie Hall. 2020. The road not taken: The sound pattern of russian and the history of contrast in phonology. Journal of Linguistics 57(2). 405–444. doi:10.1017/s0022226720000377.

Dresher, B. Elan & Daniel Currie Hall. 2022. Developments leading towards generative phonology. In B. Elan Dresher & Harry van der Hulst (eds.), The Oxford history of phonology, 372–395. Oxford: Oxford University Press.

Flier, Michael S. 1974. Lightner on Russian phonology. Russian Linguistics 1(3–4). 295–311. doi:10.1007/bf02551680.

Halle, Morris. 1951. The Old Church Slavonic conjugation: With an appendix on the Old Russian conjugation. Word 7(2). 155–167. doi:10.1080/00437956.1951.11659400.

Halle, Morris. 1959. The sound pattern of Russian: A linguistic and acoustical investigation. ’s Gravenhage: Mouton.

Halle, Morris. 1962. Phonology in generative grammar. Word 18(1–3). 54--72. doi:10.1080/00437956.1962.11659765.

Halle, Morris. 1963. O pravilakh russkogo spryazheniya (predvaritel’noe soobshchenie). American contributions to the Fifth International Congress of Slavists, Sofia, September 1963, vol. 1, 113–132. The Hague: Mouton.

Jakobson, Roman. 1929. Remarques sur l’évolution phonologique du russe comparée à celle des autres langues slaves (Travaux du Cercle linguistique de Prague 2). Prague: Jednota československých matematiků a fyziků.

Kavitskaya, Darya. 2017. Some recent developments in Slavic phonology. Journal of Slavic Linguistics 25(2). 387–414. https://www.jstor.org/stable/26535071.

Kiparsky, Paul. 2018. Formal and empirical issues in phonological typology. In Larry M. Hyman & Frans Plank (eds.), Phonological typology (Phonetics and Phonology 23), 54–106. Berlin: De Gruyter.

Lightner, Theodore M. 1963. Preliminary remarks on the morphophonemic component of Polish. Research Laboratory on Electronics quarterly progress report 71, 220–234. Massachusetts Institute of Technology.

Lightner, Theodore M. 1965. Segmental phonology of Modern Standard Russian. Cambridge, MA: Massachusetts Institue of Technology PhD thesis.

Lightner, Theodore M. 1967. On the phonology of Russian conjugation. Linguistics 5. 35–55. doi:10.1515/ling.1967.5.35.35.

Sabev, Mitko. 2023. Unstressed vowel reduction and contrast neutralisation in western and eastern Bulgarian: A current appraisal. Journal of Phonetics 99. 1012–1042. doi:10.1016/j.wocn.2023.101242.

Sériot, Patrick. 2014. Structure and the whole: East, west and non-Darwinian biology in the origins of structural linguistics. (Trans.) Amy Jacobs-Colas. Berlin: Mouton.

Shapiro, Michael. 1974. Morphophonemics as semiotic. Acta Linguistica Hafniensia 15(1). 29–49. doi:10.1080/03740463.1974.10414880.

Sussex, Roland. 1976. On the notions “underlying structure” and “process” in modern Slavic linguistics. International Journal of Slavic Linguistics and Poetics 22. 19–24.

Tilkov, Dimitŭr (ed.). 1982. Gramatika na sŭvremenniya bŭlgarskiya knizhoven ezik. Vol. 1. Sofia: Izdatelstvo na Bŭlgarskata akademiya na naukite.

Toman, Jindřich. 1995. The magic of a common language: Jakobson, Mathesius, Trubetzkoy, and the Prague Linguistic Circle (Current Studies in Linguistics 25). Cambridge, MA: MIT Press.

Trubetzkoy, Nikolai S. 1934. Das morphonologische System der russischen Sprache (Travaux Du Cercle Linguistique de Prague 5.2). Prague: Jednota československých matematiků a fyziků.

Trubetzkoy, Nikolai S. 1939. Grundzüge der Phonologie (Travaux Du Cercle Linguistique de Prague 7). Prague.

Trubetzkoy, Nikolai S. 1954. Altkirchenslavische Grammatik: Schrift-, Laut- und Formenlehre. (Ed.) Rudolf Jagoditsch. Wien: In Kommission bei Rudolf M. Rohrer.