|Sound change and alternation|
Sound change includes any processes of language change that affect pronunciation (phonetic change) or sound system structures (phonological change). Sound change can consist of the replacement of one speech sound (or, more generally, one phonetic feature) by another, the complete loss of the affected sound, or even the introduction of a new sound in a place where there previously was none. Sound changes can be environmentally conditioned, meaning that the change in question only occurs in a defined sound environment, whereas in other environments the same speech sound is not affected by the change. The term "sound change" refers to diachronic changes, or changes in a language's underlying sound system over time; "alternation", on the other hand, refers to surface changes that happen synchronically and do not change the language's underlying system (for example, the -s in the English plural can be pronounced differently depending on what sound it follows; this is a form of alternation, rather than sound change). However, since "sound change" can refer to the historical introduction of an alternation (such as post-vocalic /k/ in Tuscan, once [k], but now [h]), the label is inherently imprecise and often must be clarified as referring to phonetic change or restructuring.
Sound change is usually assumed to be regular, which means that it is expected to apply mechanically whenever its structural conditions are met, irrespective of any non-phonological factors (such as the meaning of the words affected). On the other hand, sound changes can sometimes be sporadic, affecting only one particular word or a few words, without any seeming regularity.
For regular sound changes, the term sound law is sometimes still used. This term was introduced by the Neogrammarian school in the 19th century and is commonly applied to some historically important sound changes, such as Grimm's law. While real-world sound changes often admit exceptions (for a variety of known reasons, and sometimes without one), the expectation of their regularity or "exceptionlessness" is of great heuristic value, since it allows historical linguists to define the notion of regular correspondence (see: comparative method).
Each sound change is limited in space and time. This means it functions within a specified area (within certain dialects) and during a specified period of time. For these (and other) reasons, some scholars[who?] avoid using the term "sound law" — reasoning that a law should not have spatial and temporal limitations — replacing the term with phonetic rule.
The formal notation of sound change
- A → B
- is to be read, "A changes into (or is replaced by, is reflected as, etc.) B". It goes without saying that A belongs to an older stage of the language in question, whereas B belongs to a more recent stage. The symbol ">" can be reversed:
- B ← A
- "(more recent) B derives from (older) A"
- POc. *t → Rot. f
- = "Proto-Oceanic *t is reflected as the Voiceless labiodental fricative [f] in the Rotuman language."
The two sides of such an equation indicate start and end points only, and do not imply that there are not additional intermediate stages. The example above is actually a compressed account of a sequence of changes: *t changed first into a voiceless dental fricative [θ] (like the initial consonant of English thin), which has yielded present-day [f]. This can be represented more fully as:
- t → θ → f
Unless a change operates unconditionally (in all environments), the context in which it applies must be specified:
- A → B /X__Y
- = "A changes into B when preceded by X and followed by Y."
- It. b → v /[vowel]__[vowel], which can be simplified to just
- It. b → v /V__V (where the capital V stands for any given vowel)
- = "Intervocalic [b] (inherited from Latin) became [v] in Italian" (e.g. in caballum, dēbet → cavallo 'horse', deve 'owe (3sg.)'
A second example:
- PIr. [−cont][−voi] → [+cont]/__[C][+cont]
- = "Preconsonantal voiceless non-continuants (i.e. voiceless stops) changed into corresponding voiceless continuants (fricatives) in Proto-Iranian" when immediately followed by a continuant consonant (i.e. resonants and fricatives). Examples: Proto-Indo-Iranian *pra 'forth' → Avestan fra, *trayas "three" (masc.nom.pl.)> Av. θrayō, *čatwāras "four" (masc.nom.pl.) → Av. čaθwārō, *pśaws "of a cow" (nom. *paśu) → Av. fšāoš (nom. pasu). Note that the fricativization does not occur before stops, so *sapta "seven" → Av. hapta. (However, in the variety of Iranian underlying Old Persian, fricativization occurs in all clusters, thus Old Persian hafta "seven".)
If the symbol "#" stands for a word boundary (initial or final), the notation "/__#" = "word-finally", and "/#__" = "word-initially". For example:
- Gk. [stop] → ∅ /__#
- = "Word-final stops were deleted in Greek." Which can be simplified to
- Gk. P → ∅ / __#
- where capital P stands for any plosive.
Principles of sound change
The following statements are used as heuristics in formulating sound changes as understood within the Neogrammarian model. However, for modern linguistics, they are not taken as inviolable rules; rather, they are seen as guidelines.
Sound change has no memory: Sound change does not discriminate between the sources of a sound. If a previous sound change causes X,Y → Y (features X and Y merge as Y), a new one cannot affect only an original X.
Sound change ignores grammar: A sound change can only have phonological constraints, like X → Z in unstressed syllables. For example, it cannot only affect adjectives. The only exception to this is that a sound change may or may not recognise word boundaries, even when they are not indicated by prosodic clues. Also, sound changes may be regularized in inflectional paradigms (such as verbal inflection), in which case the change is no longer phonological but morphological in nature.
Sound change is exceptionless: If a sound change can happen at a place, it will. It affects all sounds that meet the criteria for change. Apparent exceptions are possible, due to analogy and other regularization processes, or another sound change, or an unrecognized conditioning factor. This is the traditional view, expressed by the Neogrammarians. In past decades it has been shown that sound change doesn't necessarily affect all the words that in principle it could. However, when a sound change is initiated, it often expands to the whole lexicon given enough time, though not always. For example, in Spanish the fronting of the Vulgar Latin [g] (voiced velar stop) before [i e ɛ] seems to have reached every possible word it could, but the change involving voicing of word-initial Latin [k] to [g] as in colaphus → golpe and cattus → gato did not, e.g. canna → caña. See also lexical diffusion.
Sound change is unstoppable: All languages vary from place to place and time to time, and neither writing nor media prevent this change.
Terms for changes in pronunciation
There are a number of traditional terms in historical linguistics designating types of phonetic change, either by nature or result. A number of such types are often (or usually) sporadic, that is, more or less accidents that happen to a specific form. Others affect a whole phonological system. Sound changes that affect a whole phonological system are also classified according to how they affect the overall shape of the system; see phonological change.
- Assimilation: One sound becomes more like another, or (much more rarely) two sounds become more like each other. Example: in Latin the prefix *kom- becomes con- before an apical stop ([t d]) or [n]: contactus "touched", condere "to found, establish", connūbium "legal marriage". The great majority of assimilations take place between contiguous segments, and the great majority involve the earlier one becoming more like the later one (e.g. in connūbium, m- + n becomes -nn- rather than -mm-). Assimilation between contiguous segments are (diachronically speaking) exceptionless sound laws rather than sporadic, isolated changes.
- Dissimilation: The opposite of assimilation. One sound becomes less like another, or (much more rarely) two sounds become less like each other. Examples: Latin quīnque /kʷiːnkʷe/ "five" → Proto-Romance *kinkʷe (whence French cinq, Italian cinque, etc.); Old Spanish omne "man" → Spanish hombre. The great majority of dissimilations involve segments that are not contiguous, but as with assimilations, the great majority involve an earlier sound changing with reference to a later one. Dissimilation is usually a sporadic phenomenon, but Grassmann's Law (Sanskrit and Greek) is an example of a systematic dissimilation. If the change of a sequence of fricatives such that one becomes a stop is dissimilation, then such changes as Proto-Germanic *χs to /ks/ (spelled x) in English would count as a regular sound law: PGmc. *seχs "six" → Old English siex, etc.
- Metathesis: Two sounds switch places. Example: Old English thridda became Middle English third; English comfortable pronounced as if spelled comfterble. Most such changes are sporadic, but occasionally a sound law is involved, as Romance *tl → Spanish ld, thus *kapitlu, *titlu "chapter (of a cathedral)", "tittle" → Spanish cabildo, tilde. Metathesis can take place between non-contiguous segments, as Greek amélgō "I milk" → Modern Greek armégō.
- Tonogenesis: Syllables come to have distinctive pitch contours.
- Sandhi: conditioned changes that take place at word-boundaries but not elsewhere. It can be morpheme-specific, as in the loss of the vowel in the enclitic forms of English is /ɪz/, with subsequent change of /z/ to /s/ adjacent to a voiceless consonant Frank's not here /ˈfræŋksnɒtˈhɪər/. Or a small class of elements, such as the assimilation of the /ð/ of English the, this and that to a preceding /n/ (including the /n/ of and when the /d/ is elided) or /l/: all the often /ɔːllə/, in the often /ɪnnə/, and so on. As in these examples, such features are rarely indicated in standard orthography. A striking exception is Sanskrit, whose orthography reflects a wide variety of such features: thus tat "that" is written tat, tac, taj, tad, tan depending on what the first sound of the next word is. These are all assimilations, but medial sequences do not assimilate the same way.
- Haplology: The loss of a syllable when an adjacent syllable is similar or (rarely) identical. Example: Old English Englaland became Modern English England, or the common pronunciation of probably as [ˈprɒbli]. This change usually affects commonly used words. The word haplology itself is sometimes jokingly pronounced "haplogy".
- Elision, aphaeresis, syncope, and apocope: All losses of sounds. Elision is the loss of unstressed sounds, aphaeresis the loss of initial sounds, syncope is the loss of medial sounds, and apocope is the loss of final sounds. Elision examples: in the southeastern United States, unstressed schwas tend to drop, so "American" is not /əˈmɛɹəkən/ but /ˈmɚkən/. Standard English is possum ← opossum. Syncope examples: the Old French word for "state" is estat, but then the s dropped, yielding, état. Similarly the loss of /t/ in English soften, hasten, castle, etc. Apocope examples: the final -e [ə] in Middle English words was pronounced, but is only retained in spelling as silent E. In English /b/ and /ɡ/ were apocopated in final position after nasals: lamb, long /læm/ /loŋ/.
- Epenthesis (also known as anaptyxis): The introduction of a sound between two adjacent sounds. Examples: Latin humilis → English humble; in Slavic an -l- intrudes between a labial and a following yod, as *zemya "land" → Russ. zemlya (земля). Most commonly epenthesis is in the nature of a "transitional" consonant, but vowels may be epenthetic: non-standard English film in two syllables, athlete in three. Epenthesis can be regular, as when the Indo-European "tool" suffix *-tlom everywhere becomes Latin -culum (so speculum "mirror" ← *speḱtlom, pōculum "drinking cup" ← *poH3-tlom. Some scholars reserve the term epenthesis for "intrusive" vowels and use excrescence for intrusive consonants.
- Prothesis: The addition of a sound at the beginning of a word. Example: word-initial /s/ + stop clusters in Latin gained a preceding /e/ in Old Spanish and Old French; hence, the Spanish word for "state" is estado, deriving from Latin status.
- Nasalization: Vowels followed by nasal consonants are usually nasalized. If the nasal consonant is lost but the vowel retains its nasalized pronunciation, nasalization becomes phonemic, that is, distinctive. Example: French "-in" words used to be pronounced [in], but are now pronounced [ɛ̃], and the [n] is no longer pronounced (except in cases of liaison).
Examples of specific historical sound changes
- Grimm's law
- Grassmann's law
- Verner's law
- Great Vowel Shift (English)
- High German consonant shift
- Anglo-Frisian nasal spirant law
- Kluge's Law
- Dahl's Law
- Campbell, L. (2004). Historical linguistics: an introduction. The MIT Press.
- Hock, H. H. (1991). Principles of Historical Linguistics. Mouton De Gruyter.
- McDorman, R. E. (1999). Labial Instability in Sound Change. Organizational Knowledge Press.