Paul Kiparsky
This condensed review of recent trends and developments in historical linguistics proceeds from the empirical to the conceptual, from ‘what’ to ‘how’ to ‘why’. I begin with new findings about the origins, relationships, and diversity of the world’s languages, then turn to the processes and mechanisms of change as they concern practicing historical linguists, continue with efforts to ground change in the acquisition, use, and structure of language, and conclude with a look at ongoing debates concerning the explanatory division of labour between historical and theoretical linguistics and ways to unify historical and theoretical linguistics. The emphasis throughout is on current research rather than on established textbook knowledge.1
The most visible face of historical linguistics is the study of language relationships. It has been revitalised in the last few decades by a wealth of new linguistic, historical, anthropological, and genetic evidence, innovative methods of classification, and a better understanding of how languages disperse and change. New surprising relationships are being proposed, and the internal affiliations of already established families are being reassessed.2 Binary tree models have been challenged by flatter ‘bush’ or ‘rake’ models that eliminate some formerly assumed subgroupings in favour of convergence between daughter languages, and the question of the families’ homeland, date, and dispersal has been reopened: see Sidwell (this volume) on Austro-Asiatic; Ehret (2001) on Bantu; Koch (2004), and McConvell and Bowern (2011) on Australian; Donohue and Grimes (2008) and Pawley (2011) on Malayo-Polynesian; Häkkinen (2012a, 2012c) on Uralic; Babel et al. (2009) on Western Numic; and Garrett (2006) on Indo-European, among many others.
Indo-Europeanists are continuing to debate two principal models of population spread: the farming model, which assumes expansion from Anatolia over a relatively long time span, implying an early date of 7000–6500 bce for the protolanguage (Renfrew 1987, 1999, 2000, 2001), and a more rapid expansion by pastoralists from the Pontic-Caspian steppes beginning around 4500–4000 bce (Mallory 1989). Renfrew’s theory has been popular among archaeologists, but linguists have by and large preferred the pastoralist model, on the grounds that it fits the reconstructable vocabulary, that early contacts with Uralic support the more northerly homeland (Carpelan, Parpola and Koskikallio 2001), and that Greek and Indo-Iranian are linguistically and culturally too similar to have undergone as much as 5,000 years of separate development (Darden 2001; Anthony 2007). Gray and Atkinson (2003) have weighed in with an argument for Renfrew’s narrative based on applying computational cladistic methods from biology to the IE lexicon (see also Bouckaert et al. 2012, critiqued by Häkkinen 2012b). For now the question of the homeland, date, and dispersal of Proto-IE remains open.3
Understudied languages are being documented at an increasing rate, though not nearly fast enough in view of the rate of extinction. Provisional classifications based on typological or lexicostatistical criteria, or sometimes just on gut-level intuitions, are being tested by careful comparative work (e.g. Pawley 2012). Undoubtedly many as yet unsuspected affiliations remain to be discovered. Computationally enhanced lexicostatistic methods, if validated for relatively well-understood families such as Indo-European and Austronesian, may yet turn out to yield reliable methods of establishing families and subgroups (Kessler in press; Barbangon et al. 2013; Dunn, this volume).
Meanwhile more speculative efforts to join the known language families into larger macrofamilies and super-macrofamilies continue unabated. Candidates (some of them overlapping) include Almosan and the larger Amerind (Greenberg 1987), Altaic and the larger Ural-Altaic (Starostin, Dybo, and Mudrak 2003), Indo-Uralic and the larger Nostratic (Bomhard 2008), Austric (Reid 2005), Eurasiatic (Greenberg 2000), and Indo-Pacific (Greenberg 1971, criticised in Pawley 2009). Hypotheses of remote genetic relationship can never be definitively refuted, but none of the above are supported by solid linguistic evidence (Campbell and Poser 2008), nor, for the most part, do they conform with results of population genetics (Heggarty 2012). Ural-Altaic is undoubtedly a typological and areal grouping (Janhunen 2001, 2007), but not a genetic one as far as anyone has been able to show. Nostratic is the only one of these putative families that is claimed to rest on the comparative method.4 The equally famous putative macrofamily Amerind, comprising all American languages except for Na-Dene and Eskimo-Aleut (Greenberg 1987; Greenberg and Ruhlen 1992), is a priori plausible, and seems to fit well with population genetics and archaeology, so it is not surprising that many anthropologists and archaeologists have readily taken the linguistic kinship on faith. Specialists on American languages, however, consider the Amerind family unsubstantiated on linguistic grounds because it is based only on the method of mass comparison, which is clearly flawed (Campbell and Poser 2008).
For now, the Comparative Method remains the gold standard (Weiss, this volume; Hale, this volume). The persistent superstition that it does not work for unwritten languages or for certain families has been refuted again and again – famously for Algonquian in the classic work of Bloomfield (1925, 1946), continued by Haas (1958), Goddard (1979, 1990), Garrett (2004), Berman (2006), and others. More recently the applicability of the Comparative Method has been questioned for Pama-Nyungan (Australian), but the work of Hale (1966) and many scholars since then should have laid the doubts to rest (see O’Grady and Hale [2004]; Alpher [2004], and other articles in the same volume [see also Miceli, this volume]).
The Comparative Method requires first of all large amounts of reliable data. It works best when it can build on prior synchronic analysis and internal reconstruction (Joseph 2010), with as much partial reconstruction of subgroups and subfamilies as can be done first. When these foundations are in place, hypotheses about even remote genetic relationships can yield precise testable predictions. Recent novel proposals connect the Andaman Islands’ languages Jarawa and Onge to Austronesian (Blevins 2007b), Chitimacha in Louisiana to Totonacan and Mixe-Zoquean (Brown, Wichmann, and Beck, to appear), and, most surprisingly, NaDene (Athabascan-Eyak-Tlingit, a family securely established by Krauss, see Krauss 1973) to Yeniseian, whose sole surviving member is Ket, now spoken near Krasnoyarsk in south central Russia (Vajda 2010a, 2010b). Vajda has described how, struck by the resemblance between the areally unusual verb prefixes of Ket to the inner (‘conjunct’) verb prefixes of Athabascan and Eyak, he discovered that the prefix positions in the reconstructed verb complex match up, except that Yeniseian has two extra slots, and marks aspect both directly after the auxiliary and after the root, whereas Na-Dene places the corresponding morphemes (presumably ancient enclitics) only after the root; the systematic character of this discrepancy actually makes the case for ancient affinity even stronger. What is more, many of the corresponding positions are filled by similar morphemes in Yeniseic and Na-Dene. The respective verb complexes, after Vajda (2010a: 38–40), are shown aligned in (1) and (2):5
The morphological parallelism and phonological similarities among corresponding affixes are suggestive, but the most compelling evidence for actual relationship comes from those sound correspondences which can be accounted for by independently motivated regular sound changes. (3) illustrates this with the aspectual morphemes. The consonantal difference between Proto-Algonquian-Eyak perfective *-ñ and Ket -n, and Proto-Algonquian-Eyak progressive *ł- and Ket -l is paralleled in other cognate pairs:
(3)
In contrast to the Amerind hypothesis, this work adheres to the best practices of comparative historical linguistics.6 If it holds up, the relationship would be notable for the vast geographic distance that it spans. The Yeniseian homeland is believed to have been west of Lake Baikal, which implies a migration as astounding as the great voyage that brought Austronesian speakers from Borneo to Madagascar a millennium and a half ago (Dahl 1991; Kikusawa, this volume). See further Sicoli and Holton (2014) for the hypothesis of back-migration from Beringia.
Thanks to advances in archaeology, geochemical fingerprinting, and genetics (Heggarty, this volume; Pakendorf, this volume) prehistoric population movements are becoming better understood, putting contact-based explanations for language change on a firmer footing. Consequently substratum and superstratum hypotheses, long considered a somewhat disreputable last resort (Lass 1997; Honti 2007), are becoming empirically falsifiable and are being advanced more confidently (Matras and Sakel 2007; Gvozdanovic 2012; Lucas, this volume). Campbell (1997: Ch. 9) identifies 21 linguistic areas in the Americas alone; finegrained areal analyses have been worked out for South Asia (Masica 1976), Mainland Southeast Asia (Enfield 2005), the Balkans (Friedman 2006), and Amazonia (Aikhenvald 2002), among others. Even the possibility of ‘excavating’ entire layers of lost prehistoric languages through their substratum effects in attested languages has been explored for India (Witzel 1999) and northern Europe (Salmons 1992; Schrijver 2001; Aikio 2004, 2006).
Phonological contact effects come predominantly from substratum languages (languages whose speakers have switched to a dominant language). Schrijver (2009) explains several sound changes of English on the basis of a Celtic substrate, curiously enough of the Irish rather than British Celtic type. Schrijver (2011) argues that the High German Consonant Shift in the Rhineland “is the result of speakers of Gallo-Romance switching to Germanic and replacing Germanic aspirated voiceless plosives by voiceless affricates but only in the phonetic positions in which these affricates occurred in Gallo-Romance” (p. 243). Fennists are increasingly inclined to accept Posti’s old theory that certain Proto-Finnic sound changes were caused by contact with Baltic and Germanic (Kallio 2000).7
Substratum languages can not only be the source of borrowed features, but also the cause of retention of conservative features. For example, Fenno-Swedish has preserved the quantitative contrasts of Germanic under Finnish influence (Kiparsky 2008). And the area where these contrasts were retained in medieval Swedish and Norwegian coincided approximately with the then extent of Saami, a language with a similar quantitative system. Contact with Saami and Finnish, which themselves have no pitch accents, seem to have contributed to retaining the prosodic system that was the precursor of the Scandinavian two-peak pitch accent systems (Kiparsky 2012b)—a more unusual role for contact in tonogenesis than what we are familiar with from Southeast Asia and elsewhere.
Contact-induced change has become more respectable in syntax and semantics as well. Western Uralic innovations such as finite that- and Wh-clauses, SVO word order, and prepositions, are very likely due to contact with Indo-European; North Russian syntax is influenced by Finnish. Kroch, Taylor, and Ringe (1995) attribute basic word-order changes in early English to Scandinavian influence on Northern dialects. Filppula (2013) summarises evidence for the Celtic source of syntactic innovations such as it-clefting. Persistent contact on a large scale in bilingual communities can even change the whole typology of a language. Stilo (2004) insightfully treats Iranian as a ‘buffer zone’ of Arabic and Turkic traits imposed on an Indo-European base. Ross (1996, 2001) has assembled a dossier of what he calls METATYPY, his most impressive case being the massive structural borrowing by Takia (Oceanic) from Waskia (Papuan). A favourite example in the contact literature is the (often misdescribed) moribund Greek of some Cappadocian communities, with noun morphology refurbished in agglutinative style under Turkish influence (Janse 2009).8 Drawing on Andersen (1988), Ross (2003) distinguishes three sociolinguistic/demographic types of speech communities (closed, open tightknit, open looseknit) which differ in the kinds of contact-induced and endogenous linguistic change they typically undergo (see also Heine and Kuteva 2005; Trudgill 2011; Winford 2013; and Greenhill, this volume).
This section surveys the basic empirical landscape of interest to the working historical linguist, leaving some theoretical issues for section 3.
Some items are more likely to be borrowed than others, depending on their category and place in the system (Curnow 2001; Aikhenvald 2006; Matras 2009). This is important to know in assessing the evidence for genetic relationships. Comparativists value shared irregularities and exceptions as reliable indicators of common descent, for they are relatively rarely borrowed. Loanwords, on the other hand, are the most frequent type of borrowed item (for a typology, see Haspelmath and Tadmor 2009), though some languages are unreceptive to them, because of structural incompatibility or for sociolinguistic/ideological reasons. Bound morphemes are more rarely borrowed (Johanson and Robbeets 2012) – derivation more often than inflection (Gardani 2008), and number more often than case (Gardani 2012). The reason derivational morphology appears to be comparatively easily borrowed may be simply that it can be imported through loan vocabulary that contains it, and can then spread from there into the native layer of word formation. This is not strictly speaking borrowing of morphemes, of course. Dozens of suffixes such as -ess, -ee, -esque, -ette, -oid, -nik that have come into English with loanwords from Romance and other languages are now productively added to native words. Hindi-Urdu extends the suffix -dān ‘container’ from its Persian loans to native words, e.g. pīk-dān ‘spittoon’. Among grammatical markers, it is naturally the ones closer to the stem, such as number and gender, that piggy-back on loanwords in this way. The plural Persian (ultimately Arabic) ending -āt has been nativised in Urdu, so that it can now be added even to words of Indic origin, e.g. janglāt ‘forests’. Case morphemes, on the other hand, are normally imported by bilingual code-switchers in more intimate contact settings.9
The stratified morphology-phonology association posited in Lexical Phonology and Morphology and Stratal OT predicts a parallel hierarchy for phonology. Much of English stem-level (morpho)phonology has been smuggled into English in Romance loanwords. The Romance stress rule penetrated even the native vocabulary in Early Modern English (Dresher and Lahiri 2005; Dresher 2013). Finnish, with fixed word-initial stress, has a productive system of quantitative alternations that reflect the donor languages’ stress alternations, as in politiikka ‘politics’, poliitikko ‘politician’, poliittinen ‘political’ (Austerlitz 1965). The introduction of such phonological contraband with loanwords is more common than borrowing of low-level (word-level and postlexical) phonology, which typically occurs in a substratum situation.
The general point to keep in mind is that the borrowability of an item depends not only on its category (whether it is morphology, phonology, and so on) and on the contact situation but also on its place in the linguistic system, such as its regularity, productivity, and its lexical or postlexical status. For this reason contact hypotheses have to be anchored in solid grammatical analyses.
Loanword adaptation is a balancing act between phonetic approximation and phonological/ morphological integration.10 In order to preserve as much as possible of a loanword’s phonological content, phonotactically difficult sequences are dealt with preferentially by epenthesis rather than deletion, unless epenthesis would require additional repairs, in which case deletion is resorted to (Paradis and LaCharité 1997). For example, Finnish preserves word-final consonants by epenthesis, but deletes all but the innermost consonant in initial clusters, e.g. Swedish skruv [skrʉ:v] ‘screw’ → ruuvi. The reason is that epenthesis into initial clusters would produce either exceptional non-initial stress (*sekrúuvi, *eskerúuvi) or, if the stress were moved, a further distortion of the original’s phonological content (*sékruuvi, *éskeruuvi).
In time, loanwords are integrated by making their stems inflectable and morpho-phonologically regular. in the earliest literary Finnish, foreign words ending in consonants were usually left unchanged, or adopted with various added final vowels, but the modern language generally nativises them with final -i, even when they end in coronals, which are licit word-final consonants in the native vocabulary: Japani ‘Japan’, tunneli ‘tunnel’, Inkeri ‘Inger’. Although *Japan, *tunnel, *Inger are pronouncable (cf. the native hapan ‘sour’, kannel ‘Finnish harp’, penger ‘ledge’, ‘embankment’) -V stems are preferred in nativisation because they are not subject to the morphophonological alternations of -C stems.
The closer the contact, the more systematic the rendering becomes, and the more morphophonology and morphology overrides simple phonetic similarity. The way in which borrowed items are adopted can therefore provide clues to the date and degree of contact and layering of loanwords. For example, early Germanic loanwords neutralised voicing (Finnish has no voiced stops), while later ones transposed the voicing contrast into the Finnish gemination contrast, e.g. /d/ → /t/, /t/ → /tt/ (Steinitz 1964). Another ingenious pattern of phoneme substitutions which reveals extensive contact is seen in the adaptation of foreign /f/ in Finnish. Depending on the phonological context it is rendered as:11
| (4) | a | hv intervocalically within a foot:
kaffe → kahvi ‘coffee’, biff → pihvi ‘steak’, soffa → sohva ‘sofa’, giraff → kirahvi ‘giraffe’. |
| b | h before a consonant:
saffran → sahrami ‘saffron’, saft → sahti ‘table beer’. |
|
| c | v elsewhere: (1) word- and foot-initially: fiska:l → viskaali ‘prosecutor’, färg → väri ‘colour’, unifórm → únivòrmu ‘uniform’, ingefä:ra → ínkivääri ‘ginger’, and (2) after a consonant:konfékt → konvehti ‘candy’, asfalt → asvaltti ‘asphalt’. |
Heinämäki (1976) noted that f’s basic replacement is hv, which unpacks its features into two adjacent segments, and that its other replacements are regular accommodations of hv to the phonotactic constraints of Finnish. These exclude tautosyllabic clusters and foot-final *-h. So, with the nativised stress in place and final codas supported by an epenthetic vowel, hv is syllabified as a coda and onset to the extent that these constraints allow, and any unsyllabifiable consonant is dropped: konvehti, sahrami, únivòrmu.12
All empirical and theoretical work on historical phonology must come to grips with the questions in (5).13
(5) a The constraints problem: are sound changes always natural, or can they be arbitrary? Is the direction of sound change predictable?
b The regularity problem: is sound change always regular, or can it be sporadic?
c The implementation problem: is sound change abrupt or gradual? What is the role of lexical frequency in sound change?
To be convincing, the sound changes posited in historical analyses should be natural, which is not always easy to achieve (Blust 2005, and see Blevins 2007a, 2008b). Intuitions about naturalness go only so far, so they must be grounded in typology, phonetics, and phonology. Naturalness has been a central theoretical concern of Natural Phonology (for a recent overview, see Nathan and Donegan in press), later pursued also in formal generative phonology by markedness, feature geometry, and models of the phonetics-phonology interface. Interest in the typology of phonological processes grew further with Optimality Theory, whose commitment to the perceptual and articulatory grounding of constraints and to intrinsic typological predictions attracted many phonologists to phonetic issues under the heading of Laboratory Phonology. From the wealth of recent empirical and theoretical work let us mention Cho (1999), Cser (2003), Barnes (2006), Smith (2005) and Kümmel (2007 – consonant fortition and lenition), Hajek (1997), Labov (1994) and Walker (2011 – vowel shifts, harmony, metaphony), Lahiri, Riad and Jacobs (1999), Lahiri (in press – stress shifts), Hombert, Ohala, and Ewan (1979), Thurgood (2002), Riad (2003), Ratliff (in press – tonogenesis), Blevins and Garrett (2004 – metathesis), Blevins (2008b – consonant epenthesis), Blevins (2008c – overview) and Cser (in press – overview).
The other main reality check on sound changes is that they should be regular. An even stronger requirement, the Neogrammarian Hypothesis, is that sound changes are exceptionless, and conditioned only phonologically. The best argument for the Neogrammarian Hypothesis is that phonemes don’t split spontaneously. Rather, new contrasts arise when the conditioning environment of allophones is obscured by other sound changes (secondary split). As Bloomfield noted, if the hypothesis were false, then languages would have huge incoherent phonological inventories, littered with stray sounds and clusters left over from sporadic or non-phonetically conditioned sound changes at various stages of their history, e.g. laryngeals, pitch contrasts, [ç], [β], [bh], [œ], /dhgwh-/, /mn-/, /kn-/ in English.14
The Neogrammarian Hypothesis nevertheless continues to be questioned on the basis of two sorts of phenomena. The first is based on variable phonetic realisation. The second is based on word-by-word phoneme replacement. We take them up in turn.
The more common a word or phrase, the more reduced its pronunciation. The reduction can be an imperceptible phonetic effect of a few milliseconds, or neutralisation to a categorically distinct pronunciation, as in the often-cited example of English vowel syncope (Bybee 2007):
| (6) | High frequency word: | every | [Ø] |
| Mid frequency word: | memory | [Ø ~ ə] | |
| Low frequency word: | mammary | [ə] |
Speaker-based explanations for such frequency effects hold that articulatory targets become more automatised through use (Bybee 2001; Pierrehumbert 2001, 2002). Listener-based explanations say that frequent words are more predictable, so speakers can put less effort into their articulation without risk of being misunderstood (Jurafsky et al. 2001).
In addition to frequency, such variation is sensitive to morphological and phonological factors, style, social class, gender, etc. All this is entirely compatible with the Neogrammarian Hypothesis. Structured variation is not in itself sound change; it can persist for centuries and even millennia.
As a sound change, syncope dates back to Old English, where it was phonologically conditioned by stress and syllable weight, conforming perfectly to the Neogrammarian Hypothesis (Sievers-Brunner 1942: 158–159; Campbell 1983). Once we look at the Old English change itself, rather than at the synchronic variation which it bequeathed to Middle English, and which remains productive in Modern English, we see that far from falsifying the Neogrammarian Hypothesis, it strongly supports it. To test the Neogrammarian Hypothesis one needs philologically interpreted textual material from the relevant period, or sociolinguistically aware fieldwork on ongoing sound change (Labov, Rosenfelder and Fruehwald 2013). After a thousand years, a variation pattern does not necessarily look like the sound change that originally caused it.
Structure-preserving processes can yield apparent counterexamples to the Neogrammarian Hypothesis because their isolated outputs can become lexicalised. Many syncopated trisyllabic words which had no synchronic morphological analysis (marshal, parchment), or lost it (poultry, butler) are now underlying disyllables. This is still compatible with the Neogrammarian Hypothesis, for lexicalisation of reduced forms is not sound change, as has always been recognised. In transparently suffixed words, on the other hand, such as mammary, cursory, generative, temporal, cidery, buttery, cobblery, clownery, cookery, the morphology gives evidence of their medial vowel even if it is deleted. Their trisyllabic underlying form can be acquired (‘analogically restored’) even by speakers who have only heard the syncopated form, and remain subject to variable syncope indefinitely.
A related challenge to the Neogrammarian Hypothesis is lexical diffusion (Chen and Wang 1975). Its status remains controversial. Many of the instances of lexical diffusion cited in the literature are frequency effects on variable synchronic reduction processes similar to syncope, and can be explained the same way. Phillips (2001, 2006, in press) argues that there are also sound changes that conversely affect the least frequent words first, as well as sound changes that affect the members of some word class first. Importantly, these are not reduction processes, and appear to be always discrete and structure-preserving. Such word-by-word redistribution of phonemes in the lexicon of a language is what is meant by lexical diffusion in the narrower sense (Labov 1994: 542; Kiparsky 1995). It characteristically eliminates marked values of marginally contrastive features, attributable to a learners’ bias to simplify weakly entrenched contrasts in lexical representations (Bermudez-Otero 2007). In that respect, it is arguably more akin to analogical change than to sound change, albeit not necessarily to analogy of the proportional kind (Kiparsky 1995). The spread of ‘diatonic pairs’ from nouns such as tórmènt and rébel (derived from tormént, rebél) to new cases, such as áddìct from addíct, often cited as a case of lexical diffusion, is obviously ordinary proportional analogy. The spread of accent retraction in non-derived nouns like moustache, garage, massage, cocaine, which does not extend an alternation, but simply regularises the word’s stress pattern, is a non-proportional counterpart of the same analogical process. For example, after [mə’stæʃ] is replaced by [‘mʌstæʃ], its final stress need no longer be registered in its lexical entry, which simplifies the word’s lexical representation. Except for being non-proportional, it is no different from the generalisation of accent retraction to deverbal nouns like áddìct. Both instances of the change remove individual exceptions to the rule that nouns bear main stress on a heavy penult, just as the morphological regularisation of kine to cows removes an individual exception to the plural rule.
This approach to lexical diffusion fits hand in glove with a conception of analogical change as a process which eliminates arbitrary complexity from the grammar – as grammar optimisation, or simplification if you will. This represents a break with the received view inherited from the Neogrammarians, who introduced four-part proportional analogy as a way of doing both synchronic and historical morphology without having to posit anything as abstract as morphemes or grammatical rules. This project did not work out (Morpurgo Davies 1978, 1998): proportions sometimes have to be built on ‘grammatical abstractions’ rather than on actual word forms, some analogical processes are not representable by proportions at all (‘non-proportional analogy’), and even with these enrichments proportions predict impossible analogies and fail to account for some actual ones. Kuryłowicz solved some of the problems by stipulating that the terms of proportions have to be stripped of ‘redundant’ morphemes, that the effects of automatic phonological rules have to be undone, that they can be categories rather than particular items, that the left hand side must contain formes de fondation and the right hand side the corresponding formes fondées, as well as several other ‘Laws of Analogy’. It was also found that proportions can operate on distinctive feature representations (Garey 1959) and that several proportions must be allowed to interact in a single analogical change (Leed 1970). See further Albright (2008) and Garrett (2008).
The outcome of these revisions was that the proportions themselves were left with little work to do, and the substance of the theory came to reside in the constraints imposed on them. When generative research showed that language structure and use is based on rules rather than on proportions, a natural move was to try eliminating proportions as a mechanism of analogical change as well, and see it as a process that eliminates unmotivated complications. Coupled with some reasonable assumptions about the mental representation of grammar, this turned out to predict the effects of Kuryłowicz’s extensions and to unify proportional and non-proportional analogy. It also provided a plausible acquisitional mechanism for analogical change: complications are eliminated by failing to be acquired by learners at some point. Moreover, the natural morphologists’ principles of preferred morphological organisation (system-congruity, uniformity of inflection classes, avoidance of allomorphy, compositionality, and coincidence of formal marking with morphosemantic markedness, Wurzel 1989) could all be seen as characterisations of morphological simplicity, hence as targets of natural morphological change (Kiparsky 2000a).
It has been claimed that the idea that analogy is simplification or optimisation implausibly requires that learners must “first correctly acquire the target grammar (so they can evaluate its complexity), and then … replace the acquired grammar with a simpler one … despite the fact that some of the speakers in the environment do have such grammars” (Reiss 2003: 150). Actually it just requires, like all learning theories, that learners have a preference ranking of grammars and that they move from simpler grammars to more complex ones to the extent needed to home in on the target language. Change occurs when this process does not go to completion and some aspect of the target language is never acquired.
The optimisation approach does face a real obstacle, however. Changes which in the end simplify the language can pass through quite messy intermediate stages. For example, the reduction in Sanskrit of a complex and arbitrary morphological subsystem of phrasal coordination to ordinary compounding passes through some even more complicated intermediate stages before reaching its simple goal (Kiparsky 2010). Such ‘bumpy rides’ pose a challenge for the proposal that analogy is a simplifying or regularising process, and indeed for almost any causal theory of change. Many of them can be explained by the assumption that an integrated subsystem of grammar cannot be dismantled in one fell swoop, but only in minimal steps. Formally, this idea can be reconstructed in Optimality Theory by positing a modular level-ordered organisation in which morphology and phonology are interacting subsystems. Change can then be modelled as the promotion of constraints within grammatical subsystems through a series of local optima. Although this works well for many cases of long-term drift which follow a similar ‘complex → even more complex → simple’ path, there remain recalcitrant cases to be accounted for.
Grammaticalisation is morphosyntactic and semantic change that is endogenous, but which, unlike analogy, is not based on any pre-existing patterns in the language, and gives rise to new grammatical categories (for a recent survey see Narrog and Heine 2011). These categories can express functional content previously not expressed in the language, such as new tenses or moods, and they can be new formal categories, possibly for old functional content, as when postpositions turn into case endings, or fixed word order replaces morphology as the mark of grammatical relations. I’ll refer to the two kinds of grammaticalisation as ‘functional enrichment’ and ‘formal renewal’, respectively. They often go hand in hand (new form for new function), as when free pronouns turn into pronominal clitics and end up as agreement affixes.15
Like sound changes and analogical changes, grammaticalisation processes have a characteristic directionality: “lexical categories become grammatical and grammatical categories become more grammatical” (Kuryłowicz 1965; Hopper and Traugott 2003: xv). Functional enrichment depletes lexical items of their semantic and interpreted features and eventually reduces them to purely functional elements with only uninterpreted features (Roberts and Roussou 2003; van Gelderen 2011, this volume). Formal renewal results in reduced segmental content and/or tighter prosodic bonding. These generalisations amount to the famous unidirectionality hypothesis. Changes in the reverse direction obviously occur, such as the change of genitive -s from an ending to a clitic in English and continental Scandinavian. Some researchers dispute their relevance to the unidirectionality hypothesis, on the grounds that they don’t create new categories but analogically generalise already existing constructions of the language (Plank 1995; Kiparsky 2012a). Others consider them ‘degrammaticalisations’ that weaken the unidirectionality hypothesis or falsify it outright (Harris and Campbell 1995: 336–338; Joseph 2001; Newmeyer 2001; Norde 2009); see Börjars and Vincent (2011) for discussion of the issue.
The exciting thing about grammaticalisation is that it reveals the language faculty at work. Formal renewal engenders new categories that conform to cross-linguistic generalisations regardless of their source. For example, complementisers work like complementisers whether they come from prepositions, pronouns, or verbs. Functional enrichment follows particular pathways (surveyed in Heine and Kuteva 2002 and Narrog and Heine 2011) which invite semantic and pragmatic explanation. In subjectivisation processes, extensively studied by Traugott, deontic modals (such as will, must) acquire epistemic uses, as do verbs (it promises to rain), and expressions for objective facts become subjective descriptions of beliefs, attitudes, perceptions and evaluations, e.g., apparently ‘plainly, openly’ → ‘to all appearances’ → ‘as far as I can tell’, surely ‘certainly’ → ‘I should think so’, (Traugott 1989; Schwenter and Traugott 2000; Traugott and Dasher 2002). Also remarkable are the pathways to tenses (Nicolle 2012). Go-type verbs are recruited to mark intentional/agentive future, and come-type verbs for unintentional/ nonagentive future, both of which may then be generalised into ordinary futures (Hilpert 2008). Another trajectory goes from location to present tense: expressions denoting location are the source for focalised progressives (e.g. Finnish olin lukemassa kirjaa, literally ‘I was in-reading the book’), which denote a point of time, and therefore are incompatible with stative predicates, and cannot be modified by phrases denoting extent of time. These become ordinary progressives, as has happened in English. Progressives in turn become imperfectives, and these finally end up as present tenses. The denotation at each stage is a superset of the denotation of the preceding stage (Deo 2009, 2012, this volume). In a parallel trajectory, resultative markers generalise to markers of perfective aspect, perfect tense, and finally past tense (Bybee, Perkins and Pagliuca 1994; Dahl 1985, 2000; Condoravdi and Deo 2008).
A unified explanation for grammaticalisation, especially for functional enrichment, has proved a tall order. Several avenues, partly competing, partly addressing different aspects of the phenomenon, are being pursued. Pragmatic approaches were prefigured by Meillet’s idea that grammaticalisation is due to the renewal of the expressiveness of speech forms, which has an intrinsic direction because it is a constant factor in ordinary language use. Eckardt (2006, 2011) sees grammaticalisation as a type of reanalysis in which hearers/ learners who correctly grasp the intended meaning of an utterance arrive at a novel compositional derivation of it, either by redistributing the semantic content over its constituents, or by reassigning the force of a pragmatic inference of the original state to an overt linguistic element in the innovative state, first by way of conventional implicature, then as semantic content. The scenario seems plausible, but would need to be enhanced with an account of which bits of pragmatic interpretation get semanticised and why, and why grammaticalisation is unidirectional. A semantic approach is taken by Condoravdi and Deo (2008), who show that the resultative-to-past and progressive-to-present paths are characterisable as successive stages of semantic generalisation. Kiparsky (2012a) proposes that semantic generalisation and phonological depletion are instances of simplification driven by the learner’s search for the optimal grammar. Syntactic accounts were pioneered by Roberts and Roussou’s (1999, 2003) parameter resetting model, according to which a learner faced with ambiguous evidence for a parameter setting reverts to its default value, causing a lexical category to raise to a functional head position and to become reanalysed as base-generated there.16 In the same spirit, van Gelderen (2011) postulates that Late Merge drives elements to higher functional heads, either from a lexical head or from a lower functional head,17 and Feature Economy requires minimising semantic and interpretable features in the derivation, motivating the reanalysis of specifiers as heads. For example, pronouns can lose their interpretable person and number features and turn into agreement morphemes, bearing only uninterpretable features. The result is a cycle schematically represented in (7):
Recently Traugott and Trousdale (2013) have reconceived grammaticalisation as ‘constructionalisation’ within a Construction Grammar approach (see also Bergs and Diewald 2008; Trousdale 2012).
The loss and renewal of grammatical and semantic categories is a long-term cyclic process in language change (van Gelderen 2011), e.g. Latin cantabo > cantare habeo > French chanterai > je vais chanter ‘I’ll sing’. In pragmatically triggered ‘inflationary’ change, overuse of affective or emphatic elements causes their semantic ‘bleaching’. Recently well-explored cases include the renewal of hypocoristics, polite forms of address, and negation (Jespersen’s Cycle, e.g. Kiparsky and Condoravdi 2006; Schwenter 2006; Breitbarth 2009, 2013; Willis 2012; Romero 2012; Biberauer 2012). There appears to be a similar cycle on the affirmative side, which deserves study. Early English had two versions of ‘yes’ and ‘no’, depending on whether the question was affirmative or negative:18
| Affirmative answer | Negative answer | |
| Affirmative question | yea | nay |
| Negative question | yes | no |
The more forceful yes, no, historically formed from yea, nay with a strengthener, have replaced the plain affirmation and negation.
A more global long-term kind of cyclic development seems to affect a language’s predominant morphological type. The trajectory goes from isolating to agglutinating to fusional morphology, and from there back again to isolating morphology (an idea that goes back at least to Gabelentz 1901). For example, English is close to isolating, its Germanic ancestor was fusional, Indo-European was agglutinative, and internal reconstruction from Proto-IE suggests an antecedent more isolating stage.
The development of explicit models of syntax and historical corpora has made it possible to track the time course of change with new precision, bringing the actuation problem in syntax closer to a solution.19
It is widely observed that the spread of a linguistic change through a speech community follows an S-shaped curve: innovations spread slowly at first, then fast, then go slowly again to completion. But how does change spread in the grammar? The weakest hypothesis would be that innovations spread independently in each context. Many scholars have taken for granted the natural assumption that innovations originate in favouring contexts and spread more quickly in them, reaching the least favouring contexts last. But surprisingly, corpus studies of syntactic change show that “when one grammatical option replaces another with which it is in competition across a set of linguistic contexts, the rate of replacement […] is the same in all of them” (Kroch 1989: 200). This is known as the ‘Constant Rate Effect’. The natural explanation for this generalisation is that “contexts change together because they are merely surface manifestations of a single underlying change in grammar. Differences in frequency of use of a new form across contexts reflect functional and stylistic factors, which are constant across time and independent of grammar” (p. 199).
That innovations spread at the same overall rate across all contexts does not in itself tell us which contexts are favoured, still less what causes the changes. The actualisation of change is probably determined by multiple factors, of which two important ones are saliency and function (for a review of other proposed candidates, notably markedness, see de Smet 2012).
The saliency hypothesis states that actualisation is most frequent in those innovating forms that are most like the old ones (Naro 1981; Scherre and Naro 2010). The actualisation is ‘constrained by more or less superficial generalisations based on similarity to established usage’ (de Smet 2012: 604). Essentially the same effect has been long known to govern the course of morphological analogy; Hermann Paul observed that Verner’s Law consonant alternations are levelled first in those verbs where they are not accompanied by vowel alternations; e.g. Old High German *sluoh : *sluogun is levelled to sluog : sluogun ‘hit’ (3Sg./3Pl.), whereas zõh : zugun ‘drew’ retains the alternation (Kiparsky 1992).
The functional hypothesis states that actualisation begins with those contexts where the innovation confers the maximum processing advantage, normally understood as parsing efficiency (minimisation of ambiguity, reduction of the burden on short-term memory). Studies of the development of do-support in English show that it is most frequent in exactly those cases where it serves to keep the verb and its object together: in transitive questions and transitive negative declaratives, with NP objects more than with sentential objects, and more in adverbial Wh-questions and yes-no questions that in Wh-questions with fronted objects (Kroch 1989). All these asymmetries follow directly from the parsing hypothesis. For example, verb-object adjacency favours Whither dost thou take it over Whither takest thou it, but is not a factor in the choice between Whither dost thou go and Whither goest thou. Warner (2006) suggests that the parsing advantage drives the whole change, with the other contexts being carried along because do-support is a single grammatical process, which is plausible since the parsing-based preference actually hardened into a grammatical requirement in late Middle English.20
The Constant Rate Effect has to do with the frequency distribution of competing forms across contexts. It should not be taken to exclude the possibility of syntactic change in an orderly sequence of discrete steps, each constituting a possible grammar. Large-scale change of this type is called drift. An instructive miniature-scale example is the reanalysis of the nouns key and fun as adjectives (de Smet 2012: 624). The new adjective key was at first categorically restricted to attributive position (a key first step) and began to be used predicatively several decades later (this step was [absolutely] key). The new adjective fun, on the other hand, was at first categorically restricted to predicative position, and became available for attributive use some six decades later.21 De Smet’s explanation is that the reanalysed adjectives remain favoured in the positions where the reanalysis took place, “mimicking the syntactic behavior of [their] source item” (2012: 628). He posits a single global (though not necessarily abrupt) noun-to-adjective reanalysis, followed by gradual actualisation of the new adjective, beginning with the attributive environment for key and the predicative environment for fun. The idea that the change is actualised faster in environments that resemble earlier usage contexts is inconsistent with the Constant Rate Effect, and raises the further question how, if key and fun are just adjectives, they preserve the diachronic ‘memory’ of their locus of origin. But de Smet’s insight can be reconciled with the Constant Rate Effect by assuming that the nouns became adjectives in two discrete stages. Key first became an obligatorily prenominal adjective, joining the many semantically related members of that class such as main, prime, premier, principal, chief, cardinal, top, ace, crack. In the second step, it became a regular adjective available for predicative use, though still of the non-gradable class (as prime had done in earlier British usage, see the OED’s sense 1c).Fun, on the other hand, first became an obligatorily predicative adjective, such as alive, alone, unable. In the second step, it was generalised to a regular adjective. On this view, neither word underwent a single global noun to adjective reanalysis followed by actualisation in small-scale analogical increments. Both changes took place in two smaller steps (in fact the minimal discrete steps that could be encoded in the grammar without introducing new types of adjectives into it), with across-the-board actualisation at each step, consistent with the Constant Rate Effect. At the completion of the change, they quickly acquire a frequency profile typical of adjectives.22
These data suggest that syntactic reanalysis looks forward to an available attractor structure – an existing category of the language in cases of ordinary analogy, a potential category provided by UG in cases of grammaticalisation. Actualisation, on the other hand, is conservative and utilitarian. It begins timidly with minimally salient and maximally useful instances of the new structure and leaves the most radical ones for last. As key and fun show on a small scale, the Constant Rate Effect makes available a new kind of probe into syntactic structure: different actualisation profiles diagnose different grammatical analyses.
Textbooks of historical linguistics begin with the causes of linguistic change, but they don’t address the prior question of the causes of linguistic stability. Yet the fidelity of ‘normal’ language transmission is more surprising than the failures of transmission that result in ‘change’. Keenan’s (2009) dictum that things stay as they are unless acted upon by an outside force or decay (inertia) expresses a two-part mystery. The first part of the mystery is how a language can be acquired at all. It was first posed in its full generality by Chomsky, and remains the central motif of formal linguistic theory. The second part is why learners bother to acquire their ambient social and regional dialects so accurately. Saussure was perhaps the first to raise this question, and he looked to his cardinal principle of the arbitrariness of the sign for an answer. But he was not really able to explain why arbitrariness causes such accurate reproduction. In any case, Saussurean arbitrariness is only relative: there is a good reason to prefer *goed over went, and yet exceptions and seemingly useless complications like went are ubiquitous and can persist in a language for millennia. An alternative explanation for inertia is that communicative efficiency requires some degree of complexity and/or homogeneity, which together inhibit change. The problem with that is that the complexity of languages, and the linguistic homogeneity of a speech community, exceed any communicative needs: speakers of different regional and social dialects often understand each other perfectly. Perhaps the most plausible explanation, due to Klein and Perdue (1997), appeals to the function of language as a marker of social identity. Exceptions and idiosyncrasies make the best social indicators, so they are maintained in languages (except in ad hoc contact languages and artificial languages, where perfect regularity is acceptable and even desirable), and shared by all members of the community.
But that leads to a third mystery, if learners have both the capacity and the motivation to acquire an exact replica of the language around them, how is change even possible? This has been called the logical problem of language change. It is particularly acute for theories which attribute language change to imperfect learning. A prominent family of such theories posit that change arises by covert reanalysis of ambiguous structures followed by overt extension of the new structures to unambiguous cases. In the covert phase, learners acquire a new grammar (an I-language) that generates an output (an E-language) which is indistinguishable from the old one. The covert phase is followed by an actualisation phase in which the new grammar becomes detectable in recognisably novel outputs (Langacker 1977; Harris and Campbell 1995; Roberts 2007). The question then arises how, if the old grammar is indistinguishable from the innovating grammar at the covert stage, the speech community comes to converge first on one, then on the other.
One possible answer is that change does not occur by reanalysis, but in the use of language by fluent speakers, triggered by pragmatic factors, intrinsic variation, and speech errors (Bybee and Beckner, this volume). These are undoubtedly real sources of change. But there are reasons to believe that acquisition plays a role too: (1) changes involving structural discontinuities are unlikely to be initiated by mature speakers; (2) children’s evolving grammars show features which are identical with actual or possible historical innovations of the target language;23 and (3) change is accelerated in social situations where the capacity or motivation for accurate reproduction is weakened.24
Many proponents of reanalysis theories solve the logical problem of language acquisition by assuming that changes are caused by prior changes in learners’ triggering experience. Some locate the trigger in E-language (in line with Saussure’s view that change originates in parole), and assume that random fluctuations in usage may cause the frequency of a piece of primary linguistic data to drop below a hypothetical learnability threshold, after which it fails to be acquired and disappears, or causes a grammatical parameter to be reset (Lightfoot 1991: 67–68, 1999; Clark and Roberts 1993; Hale et al. 2006: 127). The appeal to hypothetical E-language precursors of I-language change is however inconsistent with the discovery that variation (or ‘grammar competition’) is systematic, and is governed by the same constraints that underlie categorical regularities of I-language (Anttila 1997 et seq.). Others seek the triggers in prior changes in I-language, which could themselves ultimately be triggered by language contact or perhaps other factors.
Reanalysis theories have been criticised recently, especially in work on syntactic change. Apart from the triggering problem, their weakness is that on a modern understanding of syntax no general syntactic reanalysis (as opposed to minor changes such as those discussed above for key and fun) can have a covert stage. The new grammar is bound to be detectable in data that is available to learners early in the course of normal acquisition. Besides, a number of standard instances of reanalysis have been questioned on empirical grounds: nominative experiencer subjects did not arise by reanalysis of the surface dative object as a nominative subject in sentences like the woman liked those words (Allen 1995, contra Harris and Campbell 1995: 63, 83), for-infinitivals did not arise by reanalysis of prepositional phrases (Garrett 2012, contra Harris and Campbell 1995: 62).25 Whitman (2012) reviews the material and concludes that there are no credible cases of ‘rebracketing’ reanalysis, which leaves ‘relabelling’, essentially grammaticalisation, as the only surviving type. But these are precisely the cases where the existence of a covert stage is least plausible. For example, the reanalysis of a relative pronoun as a complementiser or of a verb as an auxiliary can hardly take place without detectable effects in the language.
This suggests giving up the postulate that change passes through a covert reanalysis stage, and allowing the innovative grammar to be distinct from the old one from the beginning, if only in non-salient respects. This is done by bias-driven models, which attribute change to learners’ creativity rather than to simple learning failure. Change is steered by a set of prior learners’ biases which can be strong enough to override evidence to which the learner has been exposed (Garrett and Johnson 2013; Garrett, this volume; Culbertson, Smolensky and Legendre 2012). Each individual seeks to construct an optimal internalised grammar guided by certain biases. These biases may cause learners to converge on ‘wrong’ innovative grammars, which may spread when they have a sufficient advantage over the old grammar. That is why changes can be initiated and persist in spite of accessible counterevidence. In short, instead of Saussurean change-in-parole, bias-driven models adopt the Jakobsonian conception of change as an evolutionary process governed by the same principles that constrain language itself.
The basic arguments for bias-driven models of change over purely error-driven reanalysis models are the following: (1) change involves not just loss of old ones features and categories, but the creation of new ones that have no pre-existing model in the language, such as grammaticalisation. (2) Since change is not instantaneous, a causal mechanism is required for its initiation, spread, and completion across environments over many generations – in the extreme case, drift, or long-term unidirectional change. Purely error-driven models would expect Brownian motion rather than directed drift. (3) Convergence, i.e. independent parallel change in related or unrelated languages, can be explained by acquisition biases but not by the error-driven view. It is really quite remarkable how constraints like the Final-Over-Final Constraint (FOFC, Biberauer, Holmberg, and Roberts 2013) or the Person Case Constraint (PCC) (Bonet 1994; Haspelmath 2004; Rezac 2008) emerge and are maintained independently across the languages of the world.
Bias-driven models of change depend on some characterisation of relative complexity (‘markedness’). A promising new approach uses a generalisation of Anttila’s theory of variation, introduced by Riggle (2010). The r(anking)-volume of a language l under k constraints is the number of rankings that generate l, divided by k!. The r-volume of a language turns out to be a good predictor of the direction of change (including long-term drift), consistent with a straightforward learning theory according to which the learner prefers the most probable language consistent with previously encountered data.
Let us concretise these theoretical choices for phonology. Speaker-oriented theories of sound change locate its origin in the inherent variability of speech. Speakers initiate new reduced variants for rapid and easy articulation and hyperspeech variants to achieve perceptual clarity. In a feedback loop, the target of articulation shifts as speakers accommodate to their own and others’ gradually changing outputs (see Wedel 2006 for discussion of possible mechanisms behind this). In this way physical constraints on speech shape sound change and sound systems directly. Abrupt changes such as metathesis are consigned to a different mechanism involving misperception. Learner/listener-oriented theories hold that sound change originates when speakers’ acoustic signals are misparsed, either by wrongly attributing phonetic effects to the phonological computation, or vice versa (Ohala 1981, 1993, 2012; Blevins 2004, 2006, in press; Hale, Kissock, and Reiss 2006, in press). On this view, all sound change is inherently abrupt, because of the discontinuity between speaker and hearer/learner, and the physical constraints on speech constrain sound change indirectly because hearers’ misparses reflect speakers’ implicit understanding of them.
Learner/listener oriented theories posit two types of sound change, hypocorrection and hypercorrection. Hypocorrection results from the failure to undo coarticulatory effects. Ohala (1993) illustrates hypocorrection with the schematic example of the rise of phonemic nasalised vowels in (8).
(8)
| Time 1 | Time 2 | |
| production | /VN/ >[ṼN] | /VN/> [Ṽ(N)] |
| perception | [ṼN] >/VN | [ṼN] >/Ṽ(N)/ |
The nasalisation of a vowel before a nasal consonant is the result of physical constraints of the vocal tract. At time 1, a listener ‘normalises’ a perceived [ṼN] as intended /VN/. At time 2, the nasal is ‘weakly implemented’, so that the listener may perceive only [Ṽ], and reconstructs /Ṽ/ as the intended pronunciation.26
Hypercorrection results from the converse misparsing, which occurs when intended features are perceived as coarticulatory, and the results of such misparses then become a new norm. This mechanism is responsible for dissimilation, among others, and explains why dissimilation is restricted to features which are manifested over long temporal intervals (labialisation, aspiration, glottalisation, retroflexion, pharyngealisation, place of articulation), and does not apply to manner features like ‘stop’, ‘affricate’.27 Unlike what is the case in assimilation, in dissimilation, the conditioning environment can’t be lost at the same time (since the listener analyses it as the trigger of the putative assimilation process that she undoes, i.e. imputing responsibility to it for the imagined perturbation). For similar reasons, dissimilation produces no new phonemes.
As in syntax, it is often assumed that reanalysis begins with a covert phase (Andersen 1973, 2001). A new abstract representation is acquired which initially converges on a pronunciation which is perceptually indistinguishable from the old one, and the covert distinction is later enhanced and becomes perceptible. The problem again is how learners can ever converge on the first stage of the innovation. If the two pronunciations sound the same, how can they tell them apart? And what makes learners diverge systematically from the language they actually hear?
The schema in (8) avoids this problem by interposing a prior articulatory change as a causal factor – the coda nasal is first weakened, and the weakening then triggers the perceptual change that results in distinctive vowel nasalisation and complete loss of the nasal. On this variant of the story, the emergence of the phonemic nasal vowel is not simply “the result of an unintended failure of the perceptual process,” but a consequence of a prior change in pronunciation. This version of the listener-based account nevertheless differs from a conventional articulatorily driven account of phonologisation, where allophones become distinctive when their contextual conditioning is eliminated by sound changes, in that it does not equate sound change with synchronic variation; sound change emerges from variation only when a listener fails to normalise or correct it.
A weakness of perceptual theories is that they do not account very well for neogrammarian sound change. They lead to the unwanted prediction that frequent words should resist sound change because they are heard often enough to prevent misperception (just as frequent words resist analogical remodelling in virtue of being heard often). Hale (2007: 141) addresses this problem by distinguishing between two kinds of misparsing, which we can call misperception and misanalysis. Misperceptions are singular events that happen to specific items on particular occasions, and give rise to sporadic sound change. MISANALYSES are across-the-board reinterpretations which systematically attribute a feature to some other articulatory mechanism than that being used by the source, and give rise to regular sound change. Phonemes cannot split spontaneously by misperception, for learners do not posit new contrasts for the sake of individual misperceived items. So new phonemes must arise by misanalysis. But that actually aggravates the trigger problem: how do learners arrive at a systematically erroneous analysis of the language they are exposed to?
A second often noted problem for the perceptual approach is that hypercorrection and hypocorrection are formally symmetrical, so that there is no basis for the unidirectionality of sound changes. For example, there is no intrinsic explanation for the fact that consonants normally palatalise rather than depalatalise before front vowels (i.e. hypocorrection rather than hypercorrection). Ambiguity sets the stage for reanalysis, but something else must explain the directionality of the change. Assimilation and dissimilation are quite different in other ways as well: dissimilation (by hypothesis, hypercorrection) never gives rise to new phonemes, but assimilation (hypocorrection) does. Such intrinsic asymmetries are not predicted by the theory as it stands.
Garrett and Johnson (2013) go a long way towards a solution of both problems by integrating the speaker’s and the listener/learner’s role in sound change and locating intrinsic asymmetries in the speech chain itself. The core of their theory is that “the structure imposed on the phonetic input to sound change, via the directionality of phonetic variation, is a key source of the typological patterns of sound change” (p. 59). Articulatory and perceptual biases direct sound change via motor planning, aerodynamic constraints, gestural mechanics, and perceptual parsing. Each of these bias factors furnishes the basis of different types of sound change.
Serious work in historical linguistics must be built on some explicit and consistent grammatical framework, for there is no theoretically non-committal way of talking about language. Does the rise and loss of verb-second order in Germanic involve a rule? A constraint? A construction? A word order template? A functional head? Does a language become ergative by acquiring an ergative case marker? An alignment constraint? A parameter setting? All of these things? Such choices are dictated by the linguistic theory one adopts. They matter for understanding the history, and that puts historical linguists in a good position to contribute evidence bearing on these choices. In the past historical linguists tended to opt out of the debate and to regard linguistic theory with some suspicion. This is understandable in view of their long tradition, going back to a time when language was analysed primarily as an inventory of items – categories and constructions – rather than as a system of interacting rules or constraints. On top of that lies some unexamined structuralist baggage, including the idea that categories are distributionally defined by segmenting and classifying speech, which underlies the classical phoneme and item-and-arrangement morphology. The more complex dynamic view of language offered by morphophonology, generative grammar, and Optimality Theory has had major impact on historical linguistics, and forced some rethinking of change itself. The recent flourishing of historical syntax has gone hand in hand with the development of generative syntax. Generative phonology and OT have renewed historical phonology and morphophonology (Bermúdez-Otero and Hogg 2003; Bermúdez-Otero 2006, in press; see Holt 2003).
Let us illustrate these points with examples from phonology. Any causal theory of sound change has to address the connection between change and structure:
(9) a Secondary split: how and why do new phonemes arise from allophones? Under what conditions does sound change lead to restructuring of the phonological system?
b The embedding problem: how, if at all, is sound change channelled by the phonological system?
c Typology and universals: does sound change explain typological tendencies and universals? And/or do linguistic universals explain sound change?
Bermúdez-Otero (in press) shows that Stratal OT (essentially Lexical Phonology implemented in Optimality Theory) is a good platform for formulating and answering these questions. He shows in particular that the architecture of the phonological component accounts naturally for the life-cycle of phonological processes. A sound change is phonologised when an automatic coarticulation process comes under the control of a phonetic implementation rule. It may then become stabilised as a categorical postlexical rule applying across the board in a phrasal domain. The new rule can then rise from the phrase level to higher strata, first to the word level, then to the stem level, acquiring morphological conditioning and lexical exceptions on its way, and eventually exiting into the morphology and the lexicon.
The classical phoneme has turned out to be something of a straitjacket and has not been helpful for understanding the rise and merger of phonological contrasts. Stratal OT offers a perspicuous account of two types of phonologically relevant units which fall short of being full-fledged phonemes and play a crucial role in these processes: near contrasts (near mergers and incomplete neutralisation), which are distinguished in production but not in perception (Labov, Karen and Miller 1991; Yu 2007) and quasi-phonemes (Korhonen 1969: 333–335; Harris 1990; Janda 2003; Gordon in press), such as Russian /ɨ/ and English /ł/, which are perceptually salient and specified by categorical feature values but not phonologically distinctive. Classical phonemics equated contrastiveness (unpredictable distribution), a structural notion, with distinctiveness, a perceptual notion. Synchronic and diachronic evidence shows that they must be separated, and Stratal OT provides the theoretical tools for doing so. Quasi-phonemes are non-contrastive but distinctive – they are predictable but perceptually salient. Near contrasts are the fourth case: contrastive but non-distinctive (Kiparsky in press).
| contrastive | non-contrastive | |
| distinctive | phonemes | quasi-phonemes |
| non-distinctive | near contrasts | allophones |
The conjecture is that all phonemes arise as quasi-phonemes, and that all mergers pass through a near-merger stage.
The move from item-and-arrangement morphology to morphophonology is proving helpful in languages with complex word structure, such as Uralic (Nikolaeva 2000) and some of the Indo-European languages. In the latter, the study of word accentuation had been dominated by a paradigm-centred approach which reifies various accent types (acrostatic, hysterokinetic, amphikinetic, proterokinetic, etc., see Meier-Brügger 2002/2003 for a summary). The compositional approach derives the accentuation of words from the lexically specified accentual features of their constituent morphemes, together with a set of general phonological rules or constraints. It is typologically well supported (Dybo 2011; Mamet 2011), and allows a unified analysis not only of athematic primary inflection, but of words of arbitrary complexity including derivation and compounding (Garde 1976; Kiparsky and Halle 1977; Dybo 1981; Kiparsky 2010). As Garde 2011 points out, accent is fundamentally a relational property, not a ‘thing’ that can be segmented and classified like a speech sound. The locus of morphophonological variation and change are not the word accents themselves but the system which assigns them, comprising the lexically specified accentual properties of morphemes and the rules by which the accent is computed from them in the lexical phonology.
Against the structuralist/generative view that the quantitative regularities revealed by variationist work should be dealt with outside the theory of grammar, e.g. phonetics, pragmatics, or ‘performance’ (e.g. Newmeyer 2001), OT work has shown that an important class of them belong in grammar (e.g. Anttila 2007). The two key discoveries that motivate bringing variation into grammar are: (1) it is governed by the same factors that govern categorical regularities; and (2) the frequency of a variant is proportional to the number of fully ranked constraint systems that generate it. A sound change goes to completion when the relevant ranking becomes obligatory.
If we adopt OT, sound change must be constraint promotion, and the effects of the promoted constraint then necessarily depend on the rest of the phonological system. Shockingly, this calls into question the time-honoured formulation of sound changes as context-sensitive replacement processes of the form ‘A changes to B in the context C__D’. There are in fact many kinds of cases where that schema is unilluminating. For example, if a tone-bearing vowel is syncopated, the tone is displaced to the right or to the left, as dictated by the existing constraints of the language. Sound changes can also be blocked from creating surface exceptions to an existing synchronic constraint of the language. For example, Old English syncope fails to take effect just in those cases where it would create a stress lapse or clash, or a prohibited syllable structure. Technically, such conditions on sound changes can be specified as conditioning factors (C__D in the rule schema), but only at the cost of a loss of the generalisation that the conditioning factors are manifestations of active phonological constraints of the language.28
OT also accounts for diachronic ‘conspiracies’, the joint satisfaction of a single constraint, typically prosodic/metrical, by diverse segmental and suprasegmental means. The classic example is the enforcement of the CV syllable canon in early Slavic by coda deletion, metathesis, degemination, prothesis of consonants, and coalescence of C+y clusters and V+nasal rhymes. The categorical elimination of contrastive vowel length in much of West Germanic was implemented by multiple changes in each language, most importantly by open syllable lengthening and gemination (Lahiri and Dresher 1999; Page 2007). In North Germanic, 3-mora syllables were eliminated by deletion of -j-, by vowel shortening, and by the insertion of anaptyctic vowels, and prevented from arising by restrictions on processes such as syncope that could have created them.
Modern linguistics also calls into question the construal of the comparative method as an algorithm by which phoneme correspondences are sorted into sets by complementary distribution, each set is associated with a proto-phoneme, and the reconstruction proceeds bottom-up from phonology to morphology and finally to syntax. A corollary of the item-centred operational approach inherited from structuralism (particularly the American variety), it may still have a place in textbooks for pedagogical reasons, but it is as foreign to the practice of historical linguistics as synchronic discovery procedures are to grammatical analysis. In real life the data sets required for phonological reconstruction do not drop from the sky onto the linguist’s desk, but must be discovered and augmented throughout the entire process as the analysis grows. New hypotheses emerge at all levels simultaneously. For example, the prefixal cognates between Ket and Na-Dene and the phonological correspondences they support could hardly have been established independently of the morphological correspondence between the prefix patterns in (1)–(2). The tables of sound correspondences that students are asked to compile, modelled on synchronic phonemic analysis, can actually be misleading since they do not represent the chronology of sound changes and their interaction with analogical changes, which are often crucial to their formulation. A reconstruction is a theory, not a product of data-processing operations; it stands and falls with the historical explanations it provides for the individual languages.
In a reversal of the project of grounding the way language changes in its structural properties as outlined in the preceding section, one could view structure as emerging from change. There are broadly two versions of this historicist programme. Evolutionary linguistics posits a sparse faculty of language and seeks to derive typological generalisations or universals as emergent outcomes of recurrent historical processes. Another approach starts from a rich faculty of language and limits historical explanation to such properties as over-generation and frequency (Myers 2002), or to matters of phonetic ‘substance’ (Hale 2007).29 These are both nontrivial research programmes which, if successful, will shed light on the language faculty.
Evolutionary linguistics is most clearly articulated by Blevins (2004) for phonology. She allows UG to provide only features and categories, while Barnes (2006) assumes that there are no restrictions at all on possible phonological systems; their structure is wholly explained by the changes that shaped it. Greenberg attributes typological preferences to the frequencies of different change types:
In general one may expect that certain phenomena are widespread in language because the ways they can arise are frequent and their stability, once they occur, is high. A rare or non-existent phenomenon arises only by infrequently occurring changes and is unstable once it comes into existence.
(Greenberg 1978: 75)
As Harris (2008: 56) points out, “…this only sets explanation forward one step. Then we must ask why these changes are infrequent, and why this construction is unstable.” She suggests that phenomena are typologically rare if they can only arise by a relatively large number of changes, or only under certain conditions (see also Harris 2010, and cf. de Vogelaer and van der Auwera 2010 for a similar idea). Thus, even though all changes and all conditions may be equally common, the more of them have to come together, the rarer the result.
Actually parametric and OT approaches already make an analogous prediction on the synchronic level: the frequency of a ‘phenomenon’ (‘property’, ‘trait’) is proportional to the number of parameter settings or constraint rankings that must be fixed for it to be manifested in the grammar (see p. 79 above on ranking volume). Structural complexity (defined in either of these ways) is probably a better predictor of rarity than the complexity or rarity of the originating change. The reason is that complex and rare changes can produce typologically unremarkable and simple structures (for example, some of the pathways to ergative case reviewed in Kiparsky 2008), and conversely, simple and common types of change can produce typologically rare and complex structures. For example, consider the English rarissima in the Konstanz database (Plank 2006): non-zero exponent for third person agreement but zero for all other persons (no. 34), neutralisation of both case and number restricted to second person (no. 70), V2 in declarative main clauses only if the first constituent is an adverbial with strong negative force (no. 81), relative pronoun as the only target for agreement in animacy (no. 84), and a definite article formally distinct from any pronoun (no. 122). As far as I can tell none of these extraordinary features of English have particularly complex historical origins. In general, every language has thousands of ‘properties’, some of which will inevitably be rare or even unique just by statistical necessity.30
The strong evolutionary programme of deriving actual linguistic universals from change requires showing that no sequence of possible changes can subvert them, i.e. that there can be no historical path to any grammar that violates them. To my knowledge this has never been achieved for any universal, and it is hard to see how it could be. For example, it is easy to imagine banal analogical changes and grammaticalisations that would create exceptions to the inviolable generalisations about pronoun systems presented in Cysouw (2003). Some intrinsic constraints must be preventing them from taking place. In contrast, Greenberg’s and Harris’ project of finding diachronic causes for typological tendencies is quite promising (Kiparsky 2006).
A more dubious kind of historicism appeals to the past to explain apparent anomalies in the present. Kuteva and Heine (2008: 220) claim that the Bulgarian enclitic article (masa-ta ‘table-the’) contradicts ‘grammatical theories expecting demonstratives and determiners to appear in similar (or even the same) position in a syntactic structure’ (prenominal tazi masa ‘this table’), and they explain the article historically as ‘frozen’ in the former position of the demonstrative. Actually the article is placed after the first available prosodic host on its left, as is normal for a clitic. It would be surprising if the non-clitic demonstrative ever had such a prosodically governed distribution, and Kuteva and Heine give no evidence that it ever did. There is no credible historical explanation here, and the real explanatory work, historically as well as synchronically, is done by the principles of clitic placement (Dost and Gribanova 2006; Harizanov and Gribanova to appear). Kuteva and Heine also explain the Swedish double determiner system as the result of contact with Western European languages with preposed articles. The Swedish system may well have arisen through contact, but it is neither ‘peculiar’, nor violates any putative linguistic universals (Hankamer and Mikkelsen 2005).
Mithun (2011) suggests a historical explanation for the arrangement of Navajo verb prefixes in a sequence of outer (‘disjunct’) prefixes followed by a sequence of inner (ȁconjunct’) prefixes; we saw the core of the latter in (2). She proposes that the reason conjunct prefixes are closer to the stem is that they are older, and claims that this historical explanation has the advantage of also accounting for the phonological and semantic differences between the two prefix layers: the disjunct prefixes consist of whole syllables, drawn from the full phonological inventory of Navajo, with relatively concrete and specific meanings, whereas the inner prefixes are restricted to coronals, glides, and laryngeals, and have abstract and diffuse meanings. This, she says, is because the conjunct prefixes have had more time to undergo phonological reduction and semantic change.
If this leaves you unsatisfied, it is probably because you want an explanation of the chronology itself. Why were the conjunct prefixes added first? Would the reverse development have been possible, in which case the disjunct prefixes would come inside in Athabascan (and to be the only ones of Eyak)? Why are the morphemes within the two prefix complexes ordered as they are? For reasonable answers to these questions we must turn after all to morphology, morphosyntax, and semantics. Operations that determine the grammatical functions (subject, object, etc.) of a verbal predicate’s thematic roles must naturally precede operations that depend on those functions, such as agreement. In a lexicalist theory in which words are interpreted cyclically from the innermost affix outward, or in a theory where morpheme order reflects syntactic depth, valency-determining prefixes (‘classifiers’) are expected to be closer to the verb stem than valency-sensitive prefixes. Further, the reason subject agreement prefixes sit closer to the stem than object agreement prefixes, and were historically added earlier, is that languages normally have object agreement only if they also have subject agreement. This typological generalisation is grounded on the principle that agreement elements (like anaphors) are oriented towards the highest (most prominent) available controller – the subject, unless agreement with it is already discharged. The diachrony is explained by the very same principles and preferences that advocates of historical explanation want it to replace.
As for the phonology of the prefixes, the morphological organisation of the verb offers a far more illuminating explanation of them than their mere age. Jaker’s (2012) Stratal OT analysis of Dogrib, building on the earlier Athabascan work of Hargus (1988) and Rice (1989), shows that the disjunct prefixes are attached at the word level, whereas the conjunct prefixes are added at the stem level (within which Jaker further distinguishes an outer and an inner layer, the latter corresponding to the portion reproduced above in [2]). He derives the phonological difference between the levels from a reversal in prosodic parsing: when the word level is reached, left-to-right iambic feet are replaced by right-to-left trochaic feet.
In general, for a diachronic fact to count as a nontrivial explanation, it should itself follow from some principle. And this principle should be independent of the synchronic fact to be explained. Otherwise historical explanations run the risk of begging the question by presupposing what they set out to eliminate.
Research on language change is hard because it requires expertise from many areas within and outside linguistics. More than any other part of linguistics, it needs to connect with neighbouring fields such as human genetics, archaeology, cultural and physical anthropology, history, and the philologies (see, among others, the chapters by Pakendorf, Heggarty, Epps, and Mailhammer in this volume). Within linguistics, it is situated at a crossroads where almost all branches of the field meet. A historical study might draw on processing and pragmatics, morphology and corpus linguistics, sociolinguistics and syntax, phonetics and formal language theory. Such connections raise some of the deepest foundational issues in the field, and at the same time make the results exceptionally interesting and accessible to the public. The Saussurian firewall between synchrony and diachrony has been effectively breached in research practice, and it may be time to recognise that fact in the academic structure of the field as well. This would involve incorporating the historical dimension into regular syntax and phonology courses, and ultimately breaking down the conventional segregation of historical linguistics into a separate discipline within linguistics.
1 Thanks to Andrew Garrett, Alex Jaker, and John Rickford for generously sharing their expertise, and to Phil Branigan for his eagle eye. The errors are on me.
2 Uralic has a new look, with no ‘Finno-Ugric’ branch and a new place for Saami (Häkkinen 1984; Itkonen 1998; Carpelan, Parpola and Koskikallio 2001; Salminen 2002; Häkkinen 2007, 2009). Häkkinen finds that Uralic first split into an eastern Ugro-Samoyedic and western Finno-Permic branch.
3 The role of farming in dispersal is widely assumed for other language families as well but not unchallenged (Donohue and Denham 2010).
4 E.g. Illič-Svityč’s spectacular explanation of the three-way Indo-European contrast between plain velars, labialised velars, and palatals on the basis of the quality of the following vowel, preserved in Uralic (Dybo 1989). If the rest of Nostratic could be worked out to the same standards, the controversy would be over.
5 Athabascan has added an outer (‘disjunct’) layer of prefixes, absent in the other branches. In the absence of a definitive reconstruction of Na-Dene, Vajda represents Na-Dene prefixes by a generalised model of the Athabascan conjunct prefixes.
6 But see the critique by Campbell (2011) and the reply in Brown, Wichmann and Beck (to appear).
7 This would be an unusual case of superstrate influence in phonology. But it might be understood in the context of the ancient Germanic practice of making the elite of conquered populations send their children to be raised in the courts of the conquerors as hostages (gīsl, Finnish kihla), eventually returning to powerful positions in their native country as fully acculturated speakers of the rulers’ language. Moreover there is population-genetic support for prehistoric language shifts from Germanic to Finnic (Sajantila and Pääbo 1995).
8 The Cappadocian situation is usually presented in overly broad strokes, overlooking the relationship of large-scale morphological borrowing to reduced radius of communication and language death (for which see Simpson, this volume). As Dawkins (1916) takes care to point out, the Turkicised morphology was confined to a few communities (Ferték, Ulaghátsh and Semenderé) where Greek was on the verge of extinction, its use being mainly confined to women and children. In Ferték the Greek men already spoke Turkish with each other, but still understood the dialect (though ‘not very freely’) and at Ulaghátsh even women were talking Turkish to their children. The other Cappadocian dialects of Greek described by Dawkins (1916) were threatened not by Turkish, but by standard Greek, which for the most part replaced them in the Cappadocian communities that settled in Greece after 1922. It is also worth noting that even in its terminal state the Ulaghatsh dialect faithfully preserved the archaic medieval clitic system and features of the clausal syntax lost in the majority of dialects including the standard language (Condoravdi and Kiparsky 2004).
9 Under exceptional sociolinguistic situations, two languages can become completely intertwined, like chimeras in biology. Michif (Cree verbs and French nouns) was spoken by descendants of native American women married to French fur traders (Bakker 1997). Mednyj Aleut (Aleut with Russian finite verb inflection) resulted from intermarriage between Russian men and Aleutian women (Thomason and Kaufman 1988).
10 For overviews of loan phonology see Kang 2010 and Uffmann in press.
11 I give Swedish source words, in standard orthography except that vowel length and non-initial stress are marked.
12 There is no reason to believe that hv was first added in full and later reduced to accommodate the syllable structure requirements. Rather, hv is an abstract target that the language approximates as best it can. When this strategy does not produce a syllabifiable sequence, the labial is minimally modified to permit syllabification. -fs- cannot be accommodated as *vs or as *hs since neither cluster is permissible; hence tofs→ tupsu ‘tuft’.
13 ‘Sound change’ is here used in its technical sense, not in the pretheoretical sense ‘change involving sounds’ (see Kiparsky 1995; Salmons 2010; and the chapters by Garrett and Hamann in this volume).
14 Irregular sound developments can arise through spelling pronunciations (Sloos 2013), effects of word associations, and taboos (Newman 1996).
15 In my view, grammaticalisation is a type of analogical change, differing from ordinary analogy in not being exemplar-driven, or being only partially exemplar-driven. See Kiparsky (1995) for an extended argument for this view, supported there among others by changes which are intermediate between analogical and grammaticalisation. Such cases are not uncommon; see e.g. Donohue (2003) for a case where pronominal clitics are grammaticalised as a second layer of agreement affixes in a language that already has one.
16 For developments of this approach see the articles in Batllori et al. (2005) and Galves etal. (2012).
17 See also Longobardi (2001) for a formal analysis of the origin of the French preposition chez from the noun casa ‘house’.
18 “ No answereth the questyon framede by the affyrmatyue…yf a man sholde aske..is an heretyke mete to translate holy scrypture into englyshe..he muste answere nay and not no. But and yf the questyon be asked. .Is not an heretyque mete to translate holy scripture into englysh. To this questyon..he muste answere no and not nay” (Thomas More, 1532, cited from OED). The affirmative contrast is illustrated by this example from the OED: “Thynk e not he is worthy to day? ys! ys! ys! Alle we saye he is worthy to day, a! a! a!” (1450)
19 For historical syntax in general, see the chapters by Frajzyngier, van Gelderen, and Barödal in this volume.
20 Other functional advantages might include production efficiency (minimisation of speaker effort) and computational efficiency, in Minimalism syntax expressed as a preference for the shortest derivation, minimisation of long-distance dependencies by the shortest move, MERGE MOVE etc.
21 A search of the Corpus of Historical American English and the Corpus of English Novels netted as the earliest unambiguously adjectival example Them air devils threw up their hats ‘n’ stomped ‘n’ hollered powerful, es ef’t were mighty fun to see a man cut t’ pieces (in a fictional dialect narrative attributed to a ‘pure-bred Yankee’ in Irving Batcheller’s novel D’ri and I, 1901), followed by rather fun (1907) and quite fun (1908). Adjectival prenominal fun is first attested in 1959(a fun evening).
22 There is considerable variance among adjectives, and fun falls well into the range. In the COCA Corpus, 29.1 per cent of a total 134 occurrences of very fun are pre-adjectival. For a sample of other adjectives after very, the corresponding figures are: entertaining 15.0 per cent (107), amusing 23.1 per cent (69), funny 24.4 per cent (first 1,000 hits), cool 24.6 per cent (552), enjoyable 28.0 per cent (57), sweet 29.0 per cent (548), loud 40.0 per cent (392), intense 49.6 per cent (440), rough 64.0 per cent (269). Figures for rather and quite are similar. Note that examples like John is fun (to talk to) are not unambiguously adjectival, cf. John is a nuisance (to talk to).
23 For example, some languages allow pronominals with coargument antecedents except when they are quantified (bound anaphors). This system emerges spontaneously in English child language. Grimshaw and Rosen (1990) report that some children accept sentences like Is mama bear washing her? but reject Is every bear washing her? Kroch (2001) notes the spontaneous emergence in German child language of Yiddish-style syntax, with embedded clauses with V2 and Topicalisation. On the relationship between phonological acquisition and change see Cook (2006).
24 For example, when population movements or other situations reduce the amount of exposure and/or the indexical value, as in pidgins and immigrant dialects, which are known to undergo rapid simplification.
25 I conjecture that it is connected to the late Middle English innovation by which IP becomes an obligatory syntactic projection (Kiparsky 1997). At that point, to joins the modals as a nonfinite Inflhead (van Gelderen 1993). Being nonfinite, an IP whose Inflis to must form a subordinate clause. But a subordinate clause needs a complementiser. This is the function of for. Thus the rise of for-to infinitives is part of – in fact, caused by – a larger syntactic shift.
26 Ohala’s /…/are not necessarily phonemes, but ‘intended pronunciations’.
27 Ohala does note counterexamples in both directions: l:r dissimilation, and the absence of voicing dissimilation.
28 On the Stratal OT model (Bermúdez-Otero 1999; Kiparsky 2000b) constraints must interact this way if and only if they are visible at the same level (e.g. the word-level phonology, or the postlexical phonology). Rule-based generative theories of sound change could accommodate such interactions descriptively as ‘rule insertion’ (King 1973), but had no explanation for when they occurred and why.
29 For a range of view on historical explanation in phonology see the articles in Linguistic Typology (2006) and Hansson (2008).
30 The rare ‘properties’ of pidgins and artificial languages might be the absence of exceptions and the lack of certain expressive resources.
Good, Jeff (ed.) 2008. Linguistic universals and language change. Oxford: Oxford University Press.
Honeybone, Patrick and Joseph Salmons (eds) in press. The Oxford handbook of historical phonology. Oxford: Oxford University Press.
Labov, William. 1994. Principles of linguistic change. Volume 1: Internal factors. Malden, MA: Blackwell.
Luraghi, Silvia and Vit Bubenik (eds) 2013. The Bloomsbury companion to historical linguistics. London: Bloomsbury.
Roberts, Ian and Anna Roussou. 2003. Syntactic change: a formal theory of grammaticalization. Cambridge: Cambridge University Press.
Tallerman, Maggie and Kathleen R. Gibson (eds) 2012. The Oxford handbook of language evolution. Oxford: Oxford University Press.
Traugott, Elizabeth and Graeme Trousdale. 2013. Constructionalization and constructional changes. Oxford: Oxford University Press.
Aikhenvald, Alexandra. 2002. Language contact in Amazonia. Oxford: Oxford University Press.
——2006. Grammars in contact: a cross-linguistic perspective. In Alexandra Aikhenvald and R. M. W. Dixon (eds) Areal grammars in contact: a cross-linguistic typology. Oxford: Oxford University Press, 1–66.
Aikio, Ante. 2004. An essay on substrate studies and the origin of Saami. In Hyvärinen, Irma, Petri Kallio and Jarmo Korhonen (eds) Etymologie, Entlehnungen und Entwicklungen: Festschrift für Jorma Koivulehto zum 70. Geburtstag. Helsinki: Société Néophilologique, 5–34.
——2006. On Germanic-Saami contacts and Saami prehistory. Journal de la SociétéFinno-Ougrienne 91: 9–55.
Albright, Adam. 2008. Explaining universal tendencies and language particulars in analogical change. In Good (ed.), 215–230.
Allen, Cynthia. 1995. Case-marking and reanalysis: grammatical relations from Old to Early Modern English. Oxford: Clarendon Press.
Alpher, Barry. 2004. Pama-Nyungan: phonological reconstruction and status as a phylogenetic group. In Bowern and Koch (eds), 93–126.
Andersen, Henning. 1973. Abductive and deductive change. Language 49: 765–793.
——1988. Centre and periphery: adoption, diffusion and spread. In Jacek Fisiak (ed.) Historicaldialectology. Berlin: Mouton de Gruyter, 39–85.
——2001. Actualization and the (uni)directionality of change. In Henning Andersen (ed.) Actualization:linguistic change in progress. Amsterdam/Philadelphia: John Benjamins, 225–248.
Anthony, David W. 2007. The horse, the wheel and language: how Bronze-Age riders from the Eurasian steppes shaped the modern world. Princeton, N.J.: Princeton University Press.
Anttila, Arto. 1997. Deriving variation from grammar. In Frans Hinskens, Roeland van Hout and Leo Wetzels (eds), Variation, change and phonological theory. Amsterdam/Philadelphia, John Benjamins, 35–68.
——2007. Variation and optionality. In de Lacy (ed.), 519–536.
Atkinson, Quentin. 2011. Phonemic diversity supports a serial founder effect model of language expansion from Africa. Science 332: 346–349.
Austerlitz, Robert. 1965. A Finnish/Swedish phonological parallel. In Carl F. Bayerschmidt and Erik J. Friis (ed.) Scandinavian studies: essays presented to H. G. Leach. Seattle: University of Washington Press and the American-Scandinavian Foundation, 111–124.
Babel, Molly, Andrew Garrett, Michael J. Houser, and Maziar Toosarvandani. 2009. Descent and diffusion in language diversification: a study of Western Numic dialectology. [Available at: www.linguistics.berkeley.edu/~mhouser/Papers/numic_dialectology.pdf; accessed 14 October 2013]
Bakker, Peter. 1997. A language of our own: the genesis of Michif, the mixed Cree-French language of the Canadian Metis. Oxford: Oxford University Press.
Bane, Max. 2010. Deriving the structure of variation from the structure of non-variation in the English dative alternation. In Mary Byram Washburn, Katherine McKinney-Bock, Erika Varis, Ann Sawyer, and Barbara Tomaszewicz (eds) Proceedings of the 28th West Coast Conference on Formal Linguistics. Somerville, MA: Cascadilla Proceedings Project, 42–50.
Barbanjon, Francois, Steven N. Evans, Luay Nakhleh, Don Ringe and Tandy Warnow. 2013. An experimental study comparing linguistic phylogenetic reconstruction methods. Diachronica 30: 143–170.
Barnes, Jonathan. 2006. Strength and weakness at the interface: positional neutralization in phonetics and phonology. Berlin: Mouton.
Batllori, Montserrat, Maria-Lluisa Hernanz, Carme Picallo and Francesc Roca. 2005. Grammaticalization and parametric variation. Oxford: Oxford University Press.
Benediktson, Hreinn. 1982. Nordic umlaut and breaking: thirty years of research (1951–1980). Nordic Journal of Linguistics 5: 1–60.
Bergs, Alexander and Gabriele Diewald. 2008. Constructions and language change. Berlin: Mouton de Gruyter.
Berman, Howard. 2006. Studies in Blackfoot prehistory. International Journal of American Linguistics 72: 264–284.
Bermüdez-Otero, Ricardo. 1999. Constraint interaction in language change: quantity in English and Germanic. PhD dissertation, University of Manchester.
——2006. Phonological change in Optimality Theory. In Keith Brown (ed.), Encyclopedia of languageand linguistics. Volume 9. 2nd edition. Oxford: Elsevier, 497–505.
——2007. Diachronic phonology. In de Lacy (ed.), 497–517.
——in, press. Amphichronic explanation and the life cycle of phonological processes. In Honeybone and Salmons (eds).
Bermüdez-Otero, Ricardo and Richard M. Hogg. 2003. The actuation problem in Optimality Theory: phonologisation, rule inversion, and rule loss. In D. Eric Holt (ed.) Optimality Theory and language change. Dordrecht: Kluwer, 91–119.
Bermúdez-Otero, Ricardo and Graeme Trousdale. forthcoming. Cycles and continua: on unidirectionality and gradualness in language change. In Terttu Nevalainen and Elizabeth Closs Traugott (eds) Handbook on the history of English: rethinking and extending approaches and methods. Oxford: Oxford University Press.
Biberauer, Teresa. 2012. Competing reinforcements: when languages opt out of Jespersen’s Cycle. In van Kemenade and de Haas (eds), 3–30.
Biberauer, Teresa, Anders Holmberg, and Ian Roberts. 2013. A syntactic universal and its consequences. Linguistic Inquiry.
Blevins, Juliette. 2004. Evolutionary phonology. Cambridge: Cambridge University Press.
——2006. A theoretical synopsis of evolutionary phonology. Theoretical Linguistics 32: 117–166.
——2007a. The importance of typology in explaining recurrent sound patterns. Linguistic Typology 11(1): 107–113.
——2007b. A long-lost sister of Proto-Austronesian? Proto-Ongan, mother of Jarawa and Onge. Oceanic Linguistics 46: 154–198.
——2008a. Phonetic explanation without compromise. Diachronica 25: 1–19.
——2008b. Consonant epenthesis. In Good (ed.), 215–230.
——2008c. Natural and unnatural sound patterns: a pocket field guide. In Klaas Willems and Ludovic De Cuypere (eds) Naturalness and iconicity in language. Amsterdam/Philadelphia: John Benjamins, 121–148.
——in, press. Evolutionary phonology. In Honeybone and Salmons (eds).
Blevins, Juliette and Andrew Garrett. 2004. The evolution of metathesis. In Bruce Hayes, Robert Kirchner, and Donca Steriade, (eds) Phonetically based phonology. Cambridge: Cambridge University Press, 117–56.
Bloomfield, Leonard. 1925. On the sound system of Central Algonquian. Language 1: 130–156.
——1946. Algonquian. In Harry Hoijer and Cornelius Osgood (eds) Linguistic structures of NativeAmerica. New York: Viking Fund Publications in Anthropology, 85–129.
Blust, Robert. 2005. Must sound change be linguistically motivated? Diachronica 22: 219–269.
Bomhard, Allan R. 2008. Reconstructing Proto-Nostratic: comparative phonology, morphology, and vocabulary. 2 Volumes. Leiden: Brill.
Bonet, Eulalia. 1994. The person-case constraint: a morphological approach. In H. Harley and C. Philips (eds) The morphosyntax connection. MIT Working Papers in Linguistics 22: 33–52.
Börjars, Kersti and Nigel Vincent. 2011. Grammaticalization and directionality. In Narrog and Heine (eds), 163–176.
Bouckaert, Remco, Philippe Lemey, Michael Dunn, Simon J. Greenhill, Alexander V. Alekseyenko, Alexei J. Drummond, Russell D. Gray, Marc A. Suchard, Quentin D. Atkinson. 2012. Mapping the origins and expansion of the Indo-European language family. Science 337: 957–960.
Bowern, Claire. 2010. Historical linguistics in Australia: trees, networks and their implications. Philosophical Transactions of the Royal Society, B 365(1559): 3845–3854.
Bowern, Claire and Harold Koch (eds). 2004. Australian languages: classification and the comparative method. Amsterdam/Philadelphia: John Benjamins.
Breitbarth, Anne. 2009. A hybrid approach to Jespersen’s Cycle in West Germanic. Journal of Comparative Germanic Linguistics 12: 81–114.
——2013. Indefinites, negation and Jespersen’s Cycle in the history of Low German. Diachronica 30: 171–201.
Brown, Cecil H., Seren Wichmann and David Beck. to appear. Chitimacha: a Mesoamerican language in the Lower Mississippi Valley. International Journal of American Linguistics.
Bybee, Joan. 2001. Phonology and language use. Cambridge: Cambridge University Press.
——2007. Frequency of use and the organization of language. Oxford: Oxford University Press.
Bybee, Joan L., Revere D. Perkins and William Pagliuca. 1994. The evolution of grammar: tense, aspect, and modality in the languages of the world. Chicago: University of Chicago Press.
Campbell, Alistair. 1983. Old English grammar. Oxford: Clarendon Press.
Campbell, Lyle. 1997. American Indian languages. The historical linguistics of Native America. Oxford: Oxford University Press.
——2001. What’s wrong with grammaticalization? Language Sciences 23: 113–161.
——2011. Review of James Kari and Ben A. Potter (eds) The Dene-Yeniseian Connection. International Journal of American Linguistics 77: 445–451.
Campbell, Lyle and William Poser. 2008. Language classification. Cambridge: Cambridge University Press.
Carpelan, Christian, Asko Parpola and Petteri Koskikallio (eds) 2001. Early contacts between Uralic and Indo-European: linguistic and archaeological considerations. Helsinki: Suomalais-Ugrilaisen Seuran Toimituksia 242.
Chen, Matthew and William S.-Y. Wang. 1975. Sound change: actuation and implementation. Language 51: 255–281.
Cho, Young-mee Yu. 1999. Parameters of consonantal assimilation. München: Lincom.
Clark, Robin and Ian Roberts. 1993. A computational approach to language learnability and language change. Linguistic Inquiry 24: 299–345.
Condoravdi, Cleo and Ashwini Deo. 2008. Aspect shifts in Indo-Aryan. Proceedings of the 18th International Congress of Linguists. Seoul: Linguistic Society of Korea, 2057–2076.
Condoravdi, Cleo and Paul Kiparsky. 2004. Clitics and clause structure: the Late Medieval Greek system. Journal of Greek Linguistics 5: 159–183.
Cook, Eung-Do. 2006. The patterns of consonantal acquisition and change in Chipewyan (Dëne Sliné). International Journal of American Linguistics 72: 236–263.
Corbett, Greville. 2000. Number. Cambridge: Cambridge University Press.
Cser, Andras. 2003. The typology and modelling of obstruent lenition and fortition processes. Budapest: Akadémiai Kiadö.
——in, press. Types of phonological change. In Honeybone and Salmons (eds).
Culbertson, Jennifer, Paul Smolensky, Géraldine Legendre. 2012. Learning biases predict a word order universal. Cognition 122: 306–329.
Curnow, Timothy J. 2001. What can be ‘borrowed’? In Alexandra Aikhenvald and R. M. W. Dixon (eds) Areal diffusion and genetic inheritance: problems in comparative linguistics. Oxford: Oxford University Press, 412–436.
Cysouw, Michael. 2003. The paradigmatic structure of person marking. Oxford: Oxford University Press.
Dahl, Östen. 1985. Tense and aspect systems. Oxford: Blackwell.
——2000. Tense and aspect in the languages of Europe. Berlin: Mouton.
Dahl, Otto Christian. 1991. Migration from Kalimantan to Madagascar. Oslo: Aschehoug.
Darden, Bill. 2001. On the question of the Anatolian origin of Indo-Hittite. In Robert Drews (ed.) Greater Anatolia and the Indo-Hittite language family. Washington: Journal of Indo-European Studies, 184–228.
Dawkins, Richard M. 1916. Modern Greek in Asia Minor. Cambridge: Cambridge University Press.
de Lacy, Paul (ed.). 2007. The Cambridge handbook of phonology. Cambridge: Cambridge University Press.
Deo, Ashwini. 2006. Tense and aspect in Indo-Aryan languages: variation and diachrony. PhD dissertation, Stanford University.
——2009. Unifying the imperfective and the progressive: partitions as quantificational domains. Linguistics and Philosophy 32(5): 475–521.
——2012. The imperfective-perfective contrast in Middle Indic. Journal of South Asian Languages 5: 3–33.
Donohue, Mark. 2003. Agreement in the Skou language: a historical account. Oceanic Linguistics 42: 479–498.
Donohue, Mark and Tim Denham. 2010. Farming and language in island Southeast Asia. Reframing Austronesian history. Current Anthropology 51: 223–256.
Donohue, Mark and Charles E. Grimes. 2008. Yet more on the position of the languages of eastern Indonesia and East Timor. Oceanic Linguistics 47: 115–158.
Dost, Ascander and Vera Gribanova. 2006. Definiteness marking in Bulgarian. In Donald Baumer, David Montero and Michael Scanlon (eds) Proceedings of the 25th West Coast Conference on Formal Linguistics. Somerville, MA: Cascadilla Proceedings Project, 132–140.
Dresher, B. Elan. 2013. The influence of loanwords on Norwegian and English stress. Nordlyd 40(1): 55–65.
Dresher, B. Elan and Aditi Lahiri. 2005. Main stress left in early Middle English. In Michael Fortescue, Eva Skafte Jensen, Jens Erik Mogensen and Lene Schasler (eds) Historical linguistics 2003. Selected papers from the 16th International Conference on Historical Linguistics. Amsterdam/ Philadelphia: John Benjamins, 75–85.
Dunn, Michael, Simon J. Greenhill, Stephen Levinson, Russell D. Gray. 2011. Evolved structure of language shows lineage-specific trends in word-order universals. Nature 473: 79–82.
Durie, Mark and Malcolm Ross (eds). 1996. The Comparative Method reviewed: regularity and irregularity in language change. Oxford: Oxford University Press.
Dybo, Vladimir. 1981. Slavjanskaja akcentologija. Moskva: Nauka.
——1989. Indo-European and East-Nostratic velar stops. In Vitaly Shevoroshkin (ed.) Explorations inlanguage macrofamilies. Bochum: Universitätsverlag Brockmeyer, 41–46.
——2011. Značenie zapadnokavkazskoj akcentnoj sistemy dlja izučenija balto-slavjanskoj i japonskix akcentnyx sistem. In Tijmen Pronk and Rick Derksen (eds) Accent matters: papers on Balto-Slavic accentology. New York: Rodopi.
Eckardt, Regine. 2006. Meaning change in grammaticalization. Oxford: Oxford University Press.
——2011. Semantic reanalysis and language change. Language and Linguistics Compass 5: 33–46.
Ehret, Christopher. 2001. Bantu expansions: re-envisioning a central problem of early African history. The International Journal of African Historical Studies 34: 5–41.
Enfield, N. J. 2005. Areal linguistics and Mainland Southeast Asia. Annual Review of Anthropology 34: 181–206.
Evans, Nicholas and Stephen C. Levinson. 2009. The myth of language universals: language diversity and its importance for cognitive science. Behavioral and Brain Sciences 32: 429–492.
Evans, Nicholas and Patrick McConvell. 1998. The enigma of Pama-Nyungan expansion in Australia. In Roger Blench and Matthew Spriggs (eds) Archaeology and language II: Archaeological data and linguistic hypotheses. London: Routledge, 174–191.
Filppula, Markku. 2013. Contact and the early history of English. In Raymond Hickey (ed.) The handbook of language contact. Chichester: Wiley-Blackwell, 432–453.
Forchheimer, Paul. 1953. The category of person in language. Berlin: de Gruyter.
Friedman, Victor A. 2006. Balkanizing the Balkan sprachbund: a closer look at grammatical permeability and feature distribution. In Alexandra Aikhenvald and Robert M.W. Dixon (eds) Grammars in contact: a cross-linguistic typology. Oxford: Oxford University Press, 1–66.
Gabelentz, Georg von der. 1901. Die Sprachwissenschaft; ihre Aufgaben, Methoden und bisherigen Ergebnisse. Leipzig: Tauchnitz.
Galves, Charlotte, Sonia Cyrino, Ruth Lopes, Filomena Sandalo and Juanito Avelar. 2012. Parameter theory and linguistic change. Oxford: Oxford University Press.
Gardani, Francesco. 2008. Borrowing of inflectional morphemes in language contact. Frankfurt am Main: Peter Lang.
——2012. Plural across inflection and derivation, fusion and agglutination. In Lars Johanson and Martine Robbeets (eds), 71–97.
Garde, Paul. 1976. Histoiredel’accentuation slave. Paris: Institut d’études slaves.
——2011. Spécificités de l’accentologie. Baltistica Supplement 7: 87–102.
Garey, Howard. 1959. Verte, grande and longue: three types of analogy in Old French. Language 35: 605–611.
Garrett, Andrew. 2004. The evolution of Algic verbal stem structure: new evidence from Yurok. In Marc Ettlinger, Nicholas Fleisher and Mischa Park-Doob (eds) Proceedings of the Thirtieth Annual Meeting of the Berkeley Linguistics Society, February 13–16, 2004: Special Session on the Morphology of American Indian Languages. Berkeley: Berkeley Linguistics Society, 46–60.
——2006. Convergence in the formation of Indo-European subgroups: phylogeny and chronology. In Peter Forster and Colin, Renfrew (eds) Phylogenetic methods and the prehistory of languages. Cambridge: McDonald Institute for Archaeological Research, 138–151.
——2008. Paradigmatic uniformity and markedness. In Good (ed.), 125–143.
——2012. The historical syntax problem: reanalysis and directionality. In Jonas, Whitman and Garrett (eds), 52–72.
Garrett, Andrew and Keith Johnson. 2013. Phonetic bias in sound change. In Alan C. L. Yu (ed.) Origins of sound change: approaches to phonologization. Oxford: Oxford University Press, 51–97.
Gelderen, Elly van. 1993. The rise of functional categories. Amsterdam/Philadelphia: John Benjamins.
——2004. Grammaticalization as economy. Amsterdam/Philadelphia: John Benjamins.
——2011. The linguistic cycle: language change and the language faculty. Oxford: Oxford University Press.
Gell-Mann, Murray and Merritt Ruhlen. 2011. The origin and evolution of word order. Proceedings ofthe National Academy of Sciences. 108(42): 17290–17295.
Goddard, Ives. 1979. Comparative Algonquian. In Lyle Campbell and Marianne Mithun (eds) The languages of native North America: historical and comparative assessment. Austin: University of Texas Press, 70–132.
——1994. The west-to-east cline in Algonquian dialectology. In William Cowan (ed.) Papers of the25th Algonquian Conference. Ottawa: Carleton University, 187–211.
——1990. Algonquian linguistic change and reconstruction. In Philip Baldi (ed.) Linguistic change and reconstruction methodology. Berlin: Mouton, 99–114.
Good, Jeff (ed.). 2008. Linguistic universals and language change. Oxford: Oxford University Press.
Gordon, Matthew J. in press. Chain shifts, mergers and near-mergers as changes in progress. In Honeybone and Salmons (eds).
Gray, Russell D. and Quentin D. Atkinson. 2003. Language-tree divergence times support the Anatolian theory of Indo-European origin. Nature 426: 435–439.
Greenberg, Joseph H. 1971. The Indo-Pacific hypothesis. In Thomas A. Sebeok (ed.) Current Trends in Linguistics. Volume 8: Linguistics in Oceania. The Hague: Mouton, 808–871.
——1978. Diachrony, synchrony and language universals. In Joseph H. Greenberg, Charles A. Ferguson and Edith Moravcsik (eds) Universals of human language Volume 1: Method and theory. Stanford: Stanford University, 61–91.
——1987. Language in the Americas. Stanford: Stanford University Press.
——2000. Indo-European and its closest relatives: the Eurasiatic language family. Volume 1: Grammar. Stanford: Stanford University Press.
Greenberg, Joseph H. and Merritt Ruhlen. 1992. Linguistic origins of Native Americans .Scientific American 267: 94–99.
Grimshaw, Jane and C. Rosen. 1990. Obeying the Binding Theory. In Lyn Frazier and Jill de Villiers (eds) Language processing and language acquisition. Boston: Kluwer, 357–367.
Gvozdanović, Jadranka. 2012. Criteria for differentiating inherent and contact-induced changes in linguistic reconstruction. In van Kemenade and de Haas (eds), 49–86.
Haas, Mary R. 1958. Algonkian-Ritwan: The end of a controversy. International Journal of AmericanLinguistics 24: 159–173.
Hajek, John. 1997. Universals of sound change in nasalization. Oxford: Blackwell.
Häkkinen, Jaakko. 2007. Kantauralin murteutuminen vokaalivastaavuuksien valossa. MA Dissertation. University of Helsinki. [Available at: https://oa.doria.fi/handle/10024/7044; accessed 14 October 2013].
——2009. Kantauralin ajoitus ja paikannus: perustelut puntarissa. Journal de la SociétéFinno-Ougrienne 92: 9–56.
——2012a. After the protolanguage: invisible convergence, false divergence and boundary shift. Finnisch-UgrischeForschungen 61: 7–28.
——2012b. Problems in the method and interpretations of the computational phylogenetics based on linguistic data. An example of wishful thinking: Bouckaert et al. 2012. Unpublished Ms. [Available at: www.elisanet.fi/alkupera/Problems_of_phylogenetics.pdf; accessed 14 October 2013].
——2012c. Early contacts between Uralic and Yukaghir. Per Urales ad Orientem: Iter polyphonicum multilingue. Festskrift tillägnad Juha Janhunenpåhanssextioårsdagden12februari2012. [Available at: www.sgr.fi/sust/sust264/sust264_hakkinenj.pdf].
Häkkinen, Kaisa. 1984. Wäre es schon an der Zeit, den Stammbaum zu fällen? Ural-Altaische Jahrbücher Neue Folge 4: 1–24.
Hale, Kenneth L. 1966. The Paman group of the Pama-Nyungan family. Languages of the World: Indo-Pacific. Fascicle 6. Anthropological Linguistics 8: 162–197.
Hale, Mark. 2007. Historical linguistics. Theory and method. Malden, MA: Blackwell Publishing.
Hale, Mark, Madelyn Kissock and Charles Reiss. 2006. Microvariation, variation, and the features of universal grammar. Lingua 117: 645–665.
——in, press. An I-Language approach to phonologization and lexification. In Honeybone and Salmons (eds).
Hall, Robert A., Jr. 1944. Chinese Pidgin English: Grammar and texts. Journal of the American Oriental Society 64: 95–113.
Hankamer Jorge and Line, Mikkelsen. 2005. When movement must be blocked: a Reply to Embick and Noyer. Linguistic Inquiry 36: 85–125.
Hansson, Gunnar Ólafur. 2008. Diachronic explanations of sound patterns. Language and Linguistics Compass 2: 859–893.
Hargus, Sharon. 1988. The Lexical Phonology of Sekani. PhD Dissertation, University of California, Los Angeles.
Harizanov, Boris and Vera Gribanova. to appear. 2012. Inward sensitive contextual allomorphy and its conditioning factors. Proceedings of the 43rd annual meeting of the North East Linguistic Society.
Harris, Alice and Lyle Campbell. 1995. Historical linguistics in cross-linguistic perspective. Cambridge: Cambridge University Press.
Harris, Alice C. 2008. On the explanation of typologically unusual structures. In Jeff Good (ed.) Linguistic universals and language change. Oxford: Oxford University Press, 54–76.
Harris, Alice. 2010. Explaining typologically unusual structures: the role of probability. In J. Wohlgemuth and M. Cysouw (ed.) Rethinking universals: how rarities affect linguistic theory. Berlin: de Gruyter Mouton, 91–103.
Harris, John. 1990. Derived phonological contrasts. In S. Ramsaran (ed.) Studies in the pronunciation of English: a commemorative volume in Honour of A. C. Gimson. London: Routledge, 87–105.
Haspelmath, Martin. 2004. Explaining the ditransitive person-role constraint: a usage-based approach. Unpublished Ms. [Available at: www.eva.mpg.de/lingua/staff/haspelmath/pdf/2004dpr.pdf; accessed 14 October 2013].
Haspelmath, Martin and Uri Tadmor. 2009. Loanwords in the world’s languages: a comparative handbook. Berlin: de Gruyter.
Heggarty, Paul. 2012. Commentary on Chen et al.’s ‘Worldwide Analysis of Genetic and Linguistic Relationships of Human Populations’ (1995). Human Biology 84(5): 571–578.
Heinämäki, Orvokki. 1976. Papers from the Third Scandinavian Conference of Linguistics. Text Linguistics Research Group, Academy of Finland, Turku.
Heine, Bernd and Tania Kuteva. 2002. World lexicon of grammaticalization. Cambridge: Cambridge University Press.
—— 2005. Language contact and grammatical change. Cambridge: Cambridge University Press.
Hilpert, Martin. 2008. Germanic future constructions. A usage-based approach to language change. Amsterdam/Philadelphia: John Benjamins.
Hock, Hans Henrich. 1991. Principles of historical linguistics. Berlin: Walter de Gruyter.
——2010. Suprasegmental and prosodic phonological change. In Luraghi and Bubenik (eds), 106–113.
Hoenigswald, Henry M. 1965. Language change and linguistic reconstruction. Chicago: University of Chicago Press.
Holt, D. Eric. 2003. Optimality theory and language change. Dordrecht: Kluwer.
Hombert, Jean-Marie, John J. Ohala and William G. Ewan. 1979. Phonetic explanations for the development of tones. Language 55: 37–58.
Honeybone, Patrick and J. Salmons (eds). in press. Handbook of historical phonology. Oxford: Oxford University Press.
Honti, László. 2007. Syntaktische Konstruktionen fremden Ursprungs: Eine skeptische Stellungnahme. Incontri Linguistici, 30: 49–70.
Hopper, Paul J. and Elizabeth Closs Traugott. 2003. Grammaticalization. 2nd edition. Cambridge: Cambridge University Press.
Hunley, Keith, Claire Bowern and Meghan Healy. 2012. Rejection of a serial founder effects model of genetic and linguistic coevolution. Proceedings of the Royal Society B. 279(1736): 2281–2288.
Itkonen, Terho. 1998: Reflections on Pre-Uralic and Sami-Finnic protolanguage. Finnisch-Ugrische Forschungen 54: 229–266.
Iverson, Gregory K. and Joseph C. Salmons. 2009. Naturalness and the lifecycle of sound change. In Patrick Steinkrüger and Manfred Krifka (eds) On inflection: in memory of Wolfgang U. Wurzel. Berlin: Mouton de Gruyter, 89–105.
Jaker, Alessandro. 2012. Prosodic Reversal in Dogrib (Weledeh Dialect). PhD Dissertation, Stanford University.
Jakobson, Roman, Gunnar Fant and Morris Halle. 1952. Preliminaries to speech analysis. Cambridge: MIT Press.
Janda, Richard D. 2003. ‘Phonologization’ as the start of dephoneticization – or, on sound change and its aftermath: of extension, lexicalization, and morphologization. In Brian D. Joseph and Richard D. Janda (eds) The handbook of historical linguistics. Oxford: Blackwell, 401–422.
Janhunen, Juha. 2001. Indo-Uralic and Ural-Altaic: On the diachronic implications of areal typology. In Carpelan, Parpola and Koskikallio (eds), pp. 207–220.
——2007. Typological expansion in the Ural-Altaic belt. Incontri Linguistici, 30: 71–83.
Janse, Mark. 2009. Greek-Turkish language contact in Asia Minor. Hellenic Studies 17: 37–54.
Jasanoff, Jay. 2003. Plus ja change. Lachmann’s Law in Latin. In J. H. W. Penney (ed.) Indo-European papers in honour of Anna Morpurgo Davies. New York: Oxford University Press, 405–416.
Johanson, Lars and Martine Robbeets (eds) 2012. Copies versus cognates in bound morphology. Leiden: Brill.
Jonas, Dianne, John Whitman and Andrew Garrett (eds). 2012. Grammatical change: origins, nature, outcomes. Oxford: Oxford University Press.
Joseph, Brian. 2001. Is there such a thing as grammaticalization? Language Sciences 22: 163–186.
Joseph, Brian. 2010. Internal reconstruction. In Luraghi and Bubenik (eds), 52–58.
Jurafsky, Daniel, Alan Bell, Michelle Gregory and William D. Raymond. 2001. Probabilistic relations between words: evidence from reduction in lexical production. In Joan Bybee and Paul Hopper (eds) Frequency and the emergence of linguistic structure. Amsterdam/Philadelphia: John Benjamins, 229–254.
Kallio, Petri. 2000. Posti’s Superstrate Theory at the threshold of a new millennium. In Johanna Laakso (ed.). Facing Finnic. Some challenges to historical and contact linguistics. Helsinki: Suomalais-Ugrilainen Seura, 80–99.
Kang, Y. 2010. Tutorial overview: suprasegmental adaptation in loanwords. Lingua 120: 2295–2310.
Kari, James and Ben A. Potter. 2010. The Dene-Yeniseian connection. Fairbanks: Department of Anthropology and the Alaska Native Languages Center.
Kavitskaya, Darya. 2002. Compensatory lengthening: phonetics, phonology, diachrony. London: Routledge.
Keenan, Edward. 2009. Linguistic theory and the historical creation of English reflexives. In P. Crisma and G. Longobardi (eds) Historical syntax and linguistic theory. Oxford: Oxford University Press, 17–40.
Kemenade, Ans van and Nynke de Haas (eds). 2012. Historical linguistics 2009. Amsterdam/ Philadelphia: John Benjamins.
Kessler, Brett. in press. Computational and quantitative approaches to historical phonology. In Honeybone and Salmons (eds).
King, Robert D. 1973. Rule insertion. Language 49: 551–578.
Kiparsky, Paul. 1992. Analogy. In W. Bright (ed.) International encyclopedia of linguistics. New York: Oxford University Press, 59–61.
——1995. The phonological basis of sound change. In John Goldsmith (ed.) The handbook ofphonological theory. Oxford: Blackwell, 640–670. [Reprinted in Richard D. Janda and Brian D. Joseph (eds) 2003. Handbook of historical linguistics. Oxford: Blackwell, 313–342.]
——1997. The rise of positional licensing. In Ans van Kemenade and Nigel Vincent (eds) Parameters of morphosyntactic change. Oxford: Oxford University Press, 460–493.
——1998. Sievers’ Law as prosodic optimization. In Jay Jasanoff, H. Craig Melchert and Lisi Oliver (eds), Mír Curad: Studies in honor of Calvert Watkins. Innsbruck: Innsbrucker Beiträge zur Sprachwissenschaft.
——2000a. Analogy as optimization: ‘exceptions’ to Sievers’ Law in Gothic. In Aditi Lahiri (ed.) Analogy, levelling, markedness. Principles of change in phonology and morphology. Berlin: Mouton de Gruyter, 15–46.
——2000b. Opacity and cyclicity. The Linguistic Review 17: 351–367.
——2006. Amphichronic linguistics vs. Evolutionary phonology. Theoretical Linguistics 32: 217–236.
——2008. Fenno-Swedish quantity: contrast in Stratal OT. In Andrew Nevins and Bert Vaux (eds) Rules, constraints, and phonological phenomena. Oxford University Press, 185–219.
——2008. Universals constrain change; change results in typological generalizations. In Good (ed.), 23–53.
——2010. Dvandvas, blocking, and the associative: the bumpy ride from phrase to word. Language 86(2): 302–321.
——2010. Compositional vs. paradigmatic approaches to accent and Ablaut. In Stephanie W. Jamison, H. Craig Melchert and Brent Vine (eds) Proceedings of the 21st Annual UCLA Indo-European Conference. Bremen: Hempen.
——2011. Compensatory lengthening. In Charles E. Cairns and Eric Raimy (eds) Handbook of thesyllable. Leiden/Boston: Brill, 33–70.
——2012a. Grammaticalization as optimization. In Jonas, Whitman and Garrett (eds), 15–51.
——2012b. Two-stress accent systems in Swedish. Paper presented at the fifth European Conference on Tone and Intonation, Oxford, 6–8 September 2012.
——in, press. Phonologization. In Honeybone and Salmons (eds).
Kiparsky, Paul and Cleo Condoravdi. 2006. Tracking Jespersen’s Cycle. In Mark Janse, Brian D. Joseph and A. Ralli (eds) Proceedings of the 2nd International Conference of Modern Greek Dialects and Linguistic Theory. Patras: University of Patras, 172–197.
Kiparsky, Paul and Morris Halle. 1977. Towards a reconstruction of the Indo-European accent. In Larry Hyman (ed.), Studies in stress and accent. Southern California Occasional Papers in Linguistics 4: 209–238.
Klein, W. and C. Perdue. 1997. The basic variety (or: couldn’t natural languages be much simpler?). Second Language Research 13: 301–347.
Koch, Harold. 2004. A methodological history of Australian linguistic classification. In Bowern and Koch (eds), 17–66.
Korhonen, Mikko. 1969. Die Entwicklung der morphologischen Methode im Lappischen. Finnisch-Ugrische Forschungen 37: 203–262.
Krauss, Michael E. 1973. Na-Dene. In Thomas A. Sebeok (ed.) Linguisticsin North America. The Hague: Mouton, 903–978.
Krishnamurti, Bhadriraju. 1998. Regularity of sound change through lexical diffusion: a study of s > h > Ø in Gondi dialects. Language Variation and Change 10: 193–220.
Kroch, Anthony. 1989. Reflexes of grammar in patterns of language change. Language variation and change 1: 199–244.
——2001. Syntactic change. In Mark Baltin and Cris Collins (eds) The handbook of contemporary syntactic theory. Oxford: Blackwell, 699–729.
Kroch, Anthony, Ann Taylor and Donald Ringe. 1995. The Middle English verb-second constraint: a case study in language contact and language change. In Susan C. Herring, Pieter van Reenen and Lene Schesler (eds) Textual parameters in older language. Amsterdam/Philadelphia: John Benjamins.
Kümmel, Martin. 2007. Konsonantenwandel: Bausteine zu einer Typologie des Lautwandels und ihre Konsequenzen. Wiesbaden: Reichert.
Kuryłowicz, Jerzy. 1965. The evolution of grammatical categories. Diogenes 51: 55–71.
Kuteva, Tania and Bernd Heine. 2008. On the explanatory value of grammaticalization. In Good (ed.), 215–230.
Labov, William. 1994. Principles of language change, Volume 1: Internal factors. Oxford: Blackwell.
Labov, William, Mark Karen and Corey Miller. 1991. Near-mergers and the suspension of phonemic contrast. Language Variation and Change 3: 33–74.
Labov, William, Ingrid Rosenfelder and Josef Fruehwald. 2013. One hundred years of sound change in Philadelphia: Linear incrementation, reversal, and reanalysis. Language 89: 30–65.
Lahiri, Aditi. in press. Change in stress patterns. In Honeybone and Salmons (eds).
Lahiri, Aditi and B. Elan Dresher. 1999. Open syllable lengthening in West Germanic. In Harry van der Hulst (ed.) Wordprosodic systems in the languages of Europe. Berlin: de Gruyter, 678–719.
Lahiri, Aditi, Tomas Riad, and Haike Jacobs. 1999. Diachronic prosody. Language 75: 678–719.
Langacker, Ronald W. 1977. Syntactic reanalysis. In Charles N. Li (ed.) Mechanisms of syntactic change. Austin: University of Texas Press, 57–139.
Lass, Roger. 1997. Historical linguistics and language change. Cambridge: Cambridge University Press.
Leed, Richard. 1970. Distinctive features and analogy. Lingua 26: 1–24.
Li, Michelle, Stephen Matthews and Geoff P. Smith. 2005. Pidgin English texts from the Chinese English Instructor. Hong Kong Journal of Applied Linguistics 10: 79–167.
Liberman, Anatoly. 1991. Phonologization in Germanic: umlaut and vowel shifts. In Elmer H. Antonsen and Hans Hock (eds) Stxfcrxft. Studies in Germanic linguistics: select papers from the 1st and 2nd symposium on Germanic linguistics. Amsterdam/Philadelphia: John Benjamins.
Liberman, Mark. 2011. Phonemic diversity decays ‘out of Africa’? Language Log. [Available at: http://languagelog.ldc.upenn.edu/nll/?p=3090].
Lightfoot, David. 1979. Principles of diachronic syntax. Cambridge: Cambridge University Press.
——1991. How to set parameters: arguments from language change. Cambridge, MA: MIT Press.
——1999. The development of language: acquisition, change and evolution. Oxford: Blackwell.
Longobardi, Giuseppe. 2001. Formal syntax, diachronic minimalism, and etymology: The history of French chez. Linguistic Inquiry 32: 275–302.
Luraghi, Silvia and Vit Bubenik (eds). 2010. Continuum companion to historical linguistics. London/ New York: Continuum. [Republished in 2013 as The Bloomsbury companion to historical linguistics.]
Mallory, J. P. 1989. In search of the Indo-Europeans: language, archaeology and myth. London: Thames and Hudson.
Mamet, Ingo. 2011. Cupeño Stress Shift: Diachronic Perspectives. International Journal of American Linguistics 77: 247–83.
Masica, Colin P. 1976. Defining a linguistic area: South Asia. Chicago: University of Chicago Press.
Matras, Yaron. 2009. Language contact. Cambridge: Cambridge University Press.
Matras, Yaron and Jeanette Sakel. 2007. The borrowability of structural categories. In Yaron Matras and Jeanette Sakel (eds) Grammatical borrowing in cross-linguistic perspective. Berlin: Mouton, 31–74.
McConvell, Patrick and Claire Bowern. 2011. The prehistory and internal relationships of Australian languages. Language and Linguistics Compass 5: 19–32.
Meier-Brügger, Michael. 2002. Indogermanische Sprachwissenschaft. Berlin: de Gruyter.
Mithun, Marianne. 2011. Grammaticalization and explanation. In Narrog and Heine (eds), 177–192.
Morpurgo Davies, Anna. 1978. Analogy, Segmentation and the early Neogrammarians. Transactions of the Philological Society, 76(1): 36–60.
——1998. Nineteenth-century linguistics. Volume IV of G. Lepschy (ed.) History of linguistics. London: Longman.
Mühlhäusler, Peter and Rom Harré. 1990. Pronouns and people. The linguistic construction of social and personal identity. Oxford: Blackwell.
Myers, Scott. 2002. Gaps in factorial typology: the case of voicing in consonant clusters. Rutgers Optimality Archive. [Available at: http://roa.rutgers.edu/].
Nakhleh, Luay, Donald Ringe and Tandy Warnow. 2005. Perfect phylogenetic networks: a new methodology for reconstructing the evolutionary history of natural languages. Language 81: 382–420.
Naro, Anthony Julius. 1981. The social and structural dimensions of a syntactic change. Language 57: 63–98.
Narrog, Heiko and Bernd Heine (eds). 2011. The Oxford handbook of grammaticalization. Oxford: Oxford University Press.
Nathan, Geoffrey and Patricia Donegan. in press. Natural phonology and change. In Honeybone and Salmons (eds).
Newman, John. 1996. Footnotes to a history of Cantonese. In Durie and Ross (eds), 180–217.
Newmeyer, Frederick. 1988. Language form and language function. Cambridge: MIT Press.
——2001. Deconstructing grammaticalization. Language Sciences 23: 187–229.
Nicolle, Steve. 2012. Diachrony and grammaticalization. In Robert Binnick (ed.) The Oxford handbook of tense and aspect. Oxford: Oxford University Press, 370–397.
Nikolaeva, Irina. 2000. The status of schwa in the phonological description of Northern Ostyak. Finnisch-ugrische Mitteilungen 23: 121–148.
Norde, Muriel. 2009. Degrammaticalization. Oxford: Oxford University Press.
O’Grady, Geoffrey and Kenneth Hale. 2004. The coherence and distinctiveness of the Pama-Nyungan language family within the Australian linguistic phylum. In Bowern and Koch (eds), 69–92.
Ohala, John J. 1981. The listener as a source of sound change. In C. S. Masek, R. A. Hendrick and M. F. Miller (eds), Papers from theparasession on language and behavior. Chicago: Chicago Linguistic Society, 178–203.
——1993. The phonetics of sound change. In Charles Jones (ed.) Historical linguistics: Problems and perspectives. London: Longman, 237–278.
——2012. The listener as source of sound change: an update. In Maria-Josep Solé and Daniel Recasens (eds) The initiation of sound change: perception, production, and social factors. Amsterdam/ Philadelphia: John Benjamins, 21–26.
Page, B. Richard. 2007. On the irregularity of open syllable lengthening in German. In Joseph Salmons and Shannon Dubenion-Smith (eds) Historical linguistics 2005, Amsterdam/Philadelphia: John Benjamins, 337–350.
Paradis, C. and D. LaCharité. 1997. Preservation and minimality in loanword adaptation. Journal of Linguistics 33: 379–430.
Pawley, Andrew. 2009. Greenberg’s Indo-Pacific hypothesis: an assessment. In Bethwyn Evans (ed.) Discovering history through language: papers in honour of Malcolm Ross. Canberra: Pacific Linguistics, 153–180.
——2011. On the position of Bugotu and Gela in the Guadalcanal-Nggelic subgroup of Oceanic. Oceanic, Linguistics. 50: 1–24.
Andrew. 2012. How reconstructable is Proto Trans New Guinea? Problems, progress, prospects. Language and Linguistics in, Melanesia. Special Issue 2012 (Part I): 88–164.
Phillips, Betty S. 1984. Word frequency and the actuation of sound change. Language 60: 320–342.
——2001. Lexical diffusion, lexical frequency, and lexical analysis. In Joan Bybee and Paul Hopper (eds) Frequency and the emergence of linguistic structure. Amsterdam/Philadelphia: John Benjamins, 123–136.
——2006. Word frequency and lexical diffusion. Basingstoke and New York: Palgrave Macmillan.
——in, press. Lexical diffusion in historical phonology. In Honeybone and Salmons (eds).
Pierrehumbert, Janet. 2001. Exemplar dynamics: word frequency, lenition, and contrast. In Joan Bybee and Paul Hopper (eds) Frequency effects and the emergence of lexical structure. Amsterdam/Philadelphia, John Benjamins, 137–157.
——2002. Word-specific phonetics. Laboratory Phonology VII. Berlin: Mouton de Gruyter, 101–139.
——2003. Phonetic diversity, statistical learning, and acquisition of phonology. Language and Speech 46: 115–154.
Plank, Frans. 1995. Entgrammatisierung – Spiegelbild der Grammatisierung? In Norbert Boretzky, Wolfgang Dressler and Andreas Bittner (eds) Natürlichkeitstheorie und Sprachwandel. Bochum: Brockmeyer, 199–219.
——2006. Das grammatische Raritätenkabinett. Unpublished Ms. [Available at: http://typo.uni-konstanz.de/rara/intro/index.php?pt=1; accessed 15 October 2013].
Ratliff, Martha. 2013. Tonogenesis and tone change. In Honeybone and Salmons (eds).
Reid, Lawrence A. 2005. The current status of Austric: a review and evaluation of the lexical and morphosyntactic evidence. In Laurent Sagart, Roger Blench and Alicia Sanchez-Mazas (ed.) The peopling of East Asia: putting together archaeology, linguistics and genetics. London: Routledge Curzon, 132–160.
Reiss, Charles. 2003. Language change without constraint reranking. In Eric D. Holt (ed.) Optimality Theory and language change. Dordrecht: Kluwer, 143–168.
Renfrew, Colin. 1987. Archaeology and language: The puzzle of Indo-European origins. London: Cape.
——1999. Time depth, convergence theory and innovation in Proto-Indo-European. Journal of Indo-European, Studies. 27: 257–293.
——2000. 10,000 or 5000 years ago?—Questions of time depth. In Colin Renfrew, April McMahon and Larry Trask (eds) Time depth in historical linguistics. Cambridge: McDonald Institute for Archaeological Research, 413–439.
——2001. The Anatolian origins of Proto-Indo-European and the autochthony of the Hittites. In Robert Drews Greater Anatolia and the Indo-Hittite language family. Washington: Journal of Indo-European Studies, 36–63.
Rezac, Milan. 2008. The syntax of eccentric agreement: the Person Case Constraint and absolutive displacement in Basque. Natural Language and Linguistic Theory 26: 61–106.
Riad, Tomas. 2003. Diachrony of the Scandinavian accent typology. In Paula Fikkert and Haike Jacobs (eds) Development in prosodic systems. Berlin: Mouton, 91–144.
Rice, Keren. 1989. A Grammar of Slave. Berlin: Mouton.
Riggle, Jason. 2010/To Appear. The Sampling Model of Variation. In Proceedings of the 46th annual meeting of the Chicago Linguistics Society.
Roberts, Ian. 2007. Diachronic syntax. Oxford: Oxford University Press.
Roberts, Ian and Anna Roussou. 1999. A formal approach to grammaticalization. Linguistics 37: 1011–1041.
——2003. Syntactic change: a formal theory of grammaticalization. Cambridge: Cambridge University Press.
Roberts, Philip. 2012. Towards a computer model of the historical phonology and morphology of Latin. PhD. Dissertation, Oxford University.
Romero, Sergio F. 2012. A Maya Version of Jespersen’s Cycle: the diachronic evolution of negative markers in K’iche’ Maya. International Journal of American Linguistics 78: 77–96.
Ross, Malcolm. 1996. Contact induced change and the Comparative Method: cases from Papua New Guinea. In Durie and Ross (eds), 180–217.
——2001. Contact-induced change in Oceanic languages of North-west Melanesia. In Alexandra Aikhenvald and, R. M. W. Dixon (eds) Areal diffusion and genetic inheritance: problems in comparative linguistics. Oxford: Oxford University Press, 134–166.
——2003. Diagnosing prehistoric language contact. In Raymond Hickey (ed.) Motives for language change. Cambridge: Cambridge University Press, 174–198.
Ross, Malcolm and Mark Durie. 1996. Introduction. In Durie and Ross (eds), 180–217.
Sajantila, Antti and Svante Pääbo. 1995. Language replacement in Scandinavia. Nature Genetics 11: 359–360.
Salminen, Tapani. 2002. Problems in the taxonomy of the Uralic languages in the light of modern comparative studies. In Лингвистический беспредел: сборник статей к 70- летию А. И. Кузнецовой. Moscow: Издателžство Московского университета, 44–55.
Salmons, Joseph. 1992. Northwest Indo-European vocabulary and substrate phonology. In Roger Pearson (ed.) Perspectives on Indo-European language, culture and religion: studies in honor of Edgar C. Polomé. Volume 2. Journal of Indo-European Studies Monograph Series 9: 265–279.
——2010. Segmental phonological change. In Luraghi and Bubenik (eds), 89–105.
de Saussure, Ferdinand. 1914. Coursdelinguistiquegénérale. [Translated by Wade Baskin, Course in General Linguistics. New York: Philosophical Library, 1959.]
Scherre, Maria Marta Pereira and Anthony Julius Naro. 2010. Perceptual vs. grammatical constraints and social factors in subject-verb agreement in Brazilian Portuguese. Selected Papers from NWAV 38, University of Pennsylvania Working Papers in Linguistics 16(2): 165–171.
Schrijver, Peter C. H. 2001. Lost languages of Northern Europe. In Carpelan, Parpola and Koskikallio (eds), 417–425.
——2009. Celtic influence on Old English: phonological and phonetic evidence. English language and Linguistics 13: 193–211.
——2011. The High German consonant shift and language contact. In Cornelius Hasselblatt, H. Peter Houtzagers, Remco van Pareren (eds) Language contact in times of globalization. Amsterdam/New York: Rodopi.
Schuchardt, Hugo. 1885. Über die Lautgesetze: gegen die Junggrammatiker. Berlin: R. Oppenheim.
Schwenter, Scott. 2006. Fine-tuning Jespersen’s Cycle. In Betty J. Birner and Gregory Ward (eds) Drawing the boundaries of meaning. Neo-Gricean studies in pragmatics and semantics in honor of Laurence R. Horn. Amsterdam/Philadelphia: John Benjamins.
Schwenter, Scott and Elizabeth Traugott. 2000. Invoking scalarity: the development of in fact. Journal of Historical Pragmatics 1: 7–25.
Sicoli, Mark A. and Gary Holton. 2014. Linguistic phylogenies support back-migration from Beringia to Asia. PLoS ONE 9(3): e91722.
Sievers, Eduard and Karl Brunner. 1942. Altenglische grammatik. Dritte, neubearbeitete Auflage. Tübingen: Niemeyer.
Sloos, Marjoleine. 2013. Phonological grammar and frequency: An integrated approach. Groningen: Groningen Dissertations in Linguistics.
de Smet, Hendrik. 2012. The course of actualization. Language 88: 601–633.
——2013. Spreading patterns: diffusional change in the English system of complementation. Oxford: Oxford University, Press.
Smith, Jennifer L. 2005. Phonological augmentation in prominent positions. London: Routledge.
Sproat, Richard. 2010. Ancient symbols, computational linguistics, and the reviewing practices of the general science journals. Computational Linguistics 36: 585–594.
Starostin, Sergei A., Anna V. Dybo and Oleg A. Mudrak. 2003. Etymological dictionary of the Altaic languages. Leiden: Brill.
Steinitz, Wolfgang. 1964. Zur Periodisierung der alten baltischen und germanischen Lehnwörter im Finnischen. Wissenschaftliche Zeitschrift der Humboldt-Universität zu Berlin, Gesellschafts-und sprachwissenschaftliche Reihe 13(213): 335–339.
Stilo, Donald. 2004. Iranian as buffer zone between the universal typologies of Turkic and Semitic. In Éva Agnes Csatö, Bo Isaksson and Carina Jahani (eds) Linguistic convergence and areal diffusion: case studies from Iranian, Semitic and Turkic. London: Routledge Courzon, 35–63.
Thomason, Sarah Grey and Terence Kaufman. 1988. Language contact, creolization and genetic linguistics. Berkeley: University of California Press.
Thurgood, Graham. 2002. Vietnamese and tonogenesis: Revising the model and the analysis. Diachronica 19: 333–363.
Traugott, Elizabeth. 1989. On the rise of epistemic meanings in English: an example of subjectification in semantic change. Language 65: 31–55.
——1995. Subjectification in grammaticalization. In Dieter Stein and Susan Wright (eds) Subjectivity and subjectivisation. Cambridge: Cambridge University Press, 37–54.
——2010. Grammaticalization. In Luraghi and Bubenik (eds), 271–285.
Traugott, Elizabeth and Richard Dasher. 2002. Regularity in semantic change. Cambridge: Cambridge University Press.
Traugott, Elizabeth and Graeme Trousdale. 2013. Constructionalization and constructional changes. Oxford: Oxford University Press.
Trousdale, Graeme. 2012. Grammaticalization, constructions and the grammaticalization of constructions. In Kristin Davidse, Tine Breban, Lieselotte Brems and Tanja Mortelmans (eds) Grammaticalization and language change: new reflections. Amsterdam/Philadelphia: John Benjamins, 167–198.
Trudgill, Peter. 2011. Sociolinguistic typology: social determinants of linguistic complexity. Oxford: Oxford University Press.
Uffmann, Christian. in press. Loanword adaptation. In Honeybone and Salmons (eds).
Vajda, Edward J. 2010a. A Siberian link with the Na-Dene. Anthropological Papers of the University of Alaska 5 (New Series): 31–99.
——2010b. Yeniseian, Na-Dene and historical linguistics. Anthropological Papers of the University of Alaska 5 (New Series): 100–118.
de Vogelaer, G. and J. van der Auwera. 2010. When typological rara generate rarissima: the analogical extension of verbal agreement in Dutch dialects. In J. Wohlgemuth and M. Cysouw (eds) Rara and rarissima: documenting the fringes of linguistic diversity. Berlin: Mouton de Gruyter, 47–73.
Walker, Rachel. 2011. Vowel patterns in language. Cambridge: Cambridge University Press.
Warner, Anthony. 2006. Variation and the interpretation of change in periphrastic do. In Ans van Kemenade and Bettelou Los (eds) The handbook of the history of English. Blackwell 2006, 45–67.
Wedel, Andrew. 2006. Exemplar models, evolution and language change. The Linguistic Review 23: 247–274.
Whitman, John. 2012. Misparsing and syntactic reanalysis. In van Kemenade and de Haas (eds), 70–87.
Willis, David. 2012. A minimalist approach to Jespersen’s Cycle in Welsh. In Jonas, Whitman and Garrett (eds), 93–119.
Winford, Donald. 2013. Social factors in contact languages. In Yaron Matras and Peter Bakker (eds) Contact languages: a comprehensive guide. Berlin: Mouton de Gruyter, 369–425.
Witzel, Michael. 1999. Early sources for South Asian substrate languages. Mother Tongue. Special Issue: 1–70.
Wurzel, Wolfgang Ullrich. 1989. Inflectional morphology and naturalness.
Yu, Alan. 2007. Understanding near mergers: the case of morphological tone in Cantonese. Phonology 24: 187–214.