Historic Castilian shift: "o" > "ue".

Michael Zwingli

Senior Member
English - American (U.S. - New England)
Hey, there.

I am familiar with the Spanish stem-changing verbs, such as poder and contar. There are also many Spanish nouns which have experienced a similar stem mutation, e.g. puente (< Latin pons), puerta (< Latin porta), and puerto (< Latin portus). Such stem mutations with Spanish nouns represent a historical happenstance, which appears to be unique to Castilian among the Iberian languages. Regarding these nouns, the stems often revert to the original stem in derived lemmas: puerta/"door" > porteria/"lobby". What I would like to ask about this, are: (1) what is the proper term in Spanish for this historic nominal (noun) stem-changing phenomenon, (2) what is believed to have been the cause of this in Castilian Spanish, and (3) what might the relationship of this type of stem mutation in Spanish nouns and the phenomenon of stem-changing in the conjugations of some Spanish verbs?

Thank you in advance for helping me to better understand this phenomenon of stem mutation in Spanish.
 
Last edited:
  • Michael Zwingli

    Senior Member
    English - American (U.S. - New England)
    Thank you, @Cenzontle, for identifying the term that I am querying about; "diptongacion" makes alot of sense. Is "diptongacion" the proper term for both the historical stem mutation in nouns and the stem mutation in the conjugation of stem changing verbs, as well?
     
    Last edited by a moderator:

    S.V.

    Senior Member
    Español, México
    Rather, with an /ɔ/ in American law, or British ingot, you'd have an old *nɔbe trying to reach a similar /ɔ/ in Vulgar Latin, fifty generations ago. If you try comparing law-ve nove with the same vowel, and then shifting it closer to modern nueve, without closing your lips into a 'full' /u/, you can see where it comes from. Similar in pɔnte & mɔve, as if the slightest u sound came right after p & m, right after you had your lips closed (how we make our u).

    Then yes, you'd read this also changed verbs, for example in Pawlik:
    [...] la escisión de la /ɛ/ y la /ɔ/ del latín tardío afectó a las formas verbales en que estas vocales se acentuaban [...] este paradigma se extendió por analogía a las formas verbales con las /e/ y /o/ cerradas en la raíz de manera que las alternancias entre una vocal simple y un diptongo terminaron por aparecer en verbos cuya evolución fonética normal no podía conducir a ellas [...]: pensar (pienso en vez de *penso) y mostrar (muestro en vez de *mostro).
     

    Michael Zwingli

    Senior Member
    English - American (U.S. - New England)
    If you try comparing law-ve nove with the same vowel, and then shifting it closer to modern nueve, without closing your lips into a 'full' /u/, you can see where it comes from. Similar in pɔnte & mɔve, as if the slightest u sound came right after p & m, right after you had your lips closed (how we make our u).
    Thank you, @S.V., you have given me a fair idea of how this originated in the particular pronunciation of vulgar Latin /o/ and /e/ (certainly the two most important vowels of Indo-European linguistic history) common in the central region of the Iberian Peninsula. Additionally, your comments seem to clarify why "diptongacion" occurred mainly in nouns and verbs with certain initial consonants...mainly with bilabials.
     

    Dymn

    Senior Member
    Sound changes don't usually distinguish parts of speech. You have it in nouns, verbs, adjectives (bueno), adverbs (fuera), and so on.

    Vulgar Latin had 7 vowel phonemes out of which /ɛ ɔ/ diphtongized into /je we/ in Spanish, giving rise to the 5-vowel Spanish system. This happened in stressed syllables (there was no distinction between mid-open and mid-close vowels in unstressed position), in all contexts, except before palatals (hoy, ojo, noche), and some other exceptions.

    This diphthongization also occurred in Astur-Leonese and Aragonese, with some variations, because in those languages it did happen before palatals (Asturian güey, güeyo, nueche, Aragonese güe, güello, nueit). In some Astur-Leonese varieties, /ɔ/ became /wo/ instead, including Mirandese (in Portugal): puorta, fuogo.

    This diphthongization also happened in French and Italian, but only in open syllables (syllables which don't end in a consonant): French feu, pied, but porte, sept; Italian fuoco, piede, but porta, sette. Feu is now a monophthong but it used to be /wɛ/ if I recall correctly.

    What I don't know is whether these phenomena are related or happened as a coincidence. There is no geographical continuity between Spanish and French since Catalan and Occitan only diphthongize /ɔ/ before a palatal: Catalan avui, ull, nit (Old Cat. nuit), Occitan uèi, uèlh, nuèit. I don't know about Franco-Provençal or Northern Italian languages.
     
    Last edited:

    bearded

    Senior Member
    In Castilian, the expression ''hasta luego'' often sounds like ''ata logo'', i.e. the second word has no diptongación in colloquial.
    Does it mean that in some cases the diphthongization did not completely prevail/assert itself in colloquial? Are there other similar cases?
     
    Last edited:

    Dymn

    Senior Member
    I think it may sound like logo (or something close) in rapid speech, but the underlying form is still /ˈlwego/. Diphthongization is as exactly as common in the colloquial as in the standard.

    Someone once drew attention to the fact that "el nuevo presidente" may also sound like "el novo presidente", and upon repeating it to myself various times it seems true. You can hear it in newscasts. Most likely a feature of rapid, natural speech. Hopefully someone will give a more precise answer.

    Also while aspirantization of coda s is quite common, I don't think it leads to complete deletion, it's still audibly different from ata.
     

    Swatters

    Senior Member
    French - Belgium, some Wallo-Picard
    This diphthongization also occurred in Astur-Leonese and Aragonese, with some variations, because in those languages it did happen before palatals (Asturian güey, güeyo, nueche, Aragonese güe, güello, nueit). In some Astur-Leonese varieties, /ɔ/ became /wo/ instead, including Mirandese (in Portugal): puorta, fuogo.

    This diphthongization also happened in French and Italian, but only in open syllables (syllables which don't end in a consonant): French feu, pied, but porte, sept; Italian fuoco, piede, but porta, sette. Feu is now a monophthong but it used to be /wɛ/ if I recall correctly.

    What I don't know is whether these phenomena are related or happened as a coincidence. There is no geographical continuity between Spanish and French since Catalan and Occitan only diphthongize /ɔ/ before a palatal: Catalan avui, ull, nit (Old Cat. nuit), Occitan uèi, uèlh, nuèit. I don't know about Franco-Provençal or Northern Italian languages.
    Chapter 38 (Diphthongization, by Martin Maiden) of the Oxford Guide to the Romance languages has this to say about the matter:
    Stressed open syllable diphthongization (SOSD) characterizes Oïl dialects (northern Gallo-Romance), most dialects of northern Italy (Piedmont, Liguria, much of Lombardy, Emilia, and Trentino) and Florentine (hence also Italian), together with most of western Tuscany and parts of Umbria (Perugia, Gubbio).

    General stressed syllable diphthongization (GSSD) is found in Castilian, possibly eastern Catalan (see below), northeastern Gallo-Romance (Wallon), Rovigotto (northeastern Italy), Friulian, and Vegliote. In Romanian it affects original /ɛ/ but not back vowels.

    Metaphonic diphthongization (MD) is best represented today in central and southern Italy, but recurs in Alpine Piedmontese and Lombard, and in Romagna.

    By metaphonic diphthongization, they mean that it happened (at least in the examples they provide) in both open and closed stressed syllable, but that it was condition by vowel harmony: (Ischia Neapolitan) *por'kɛllu and *por'kɛlli > pur'tʃjeddə, while *por'kɛlla and *por'kɛlle > pur'tʃeddə.

    The rest of the chapter discusses what the link between those diphthongizations are, and whether the source is ultimately stressed syllable lengthening (which works well for the SOSD type, but less so for the others) or metaphony.

    They don't discuss it in the chapter, but at least for Walloon the distribution of diphthonged closed syllable ɛ and ɔ is more complex than what they present: in Western dialects like mine only ɛ diphtongize (bɛsta > bjɛs, fenestram > fɛrnjɛs, *perdutum > pjerdy, similare > *sɛmilare > *sjɛmler > ʃɛne), but ɔ doesn't (porta > pɔrt, scortea > (ɛ)skɔrʃ, costa > kɔs), while in central and eastern varieties, ɔ does (porta > (Namur) pwart, (Liège) pwɛt; scortea > N ʃwarʃ, L hwɛs; costa > N kwas, L kwɛs). This diphthongisation isn't uniform either: with a few rare exceptions, it only happens before sCV and rCV clusters, so bellam yield bɛl and not *bjɛl, contra becomes kõt and not kwɛ̃t, etc.

    Finally, the outcome of ɛ and ɔ in open syllable differ from those in closed syllables: ferum (proud) > fiːr while ferrum (iron, one of the rare exceptions I mentioned above) > fjeːr. Compare Spanish ferum > fjero; ferrum > fjero.

    So at least for Walloon, there's a good argument to be made that it was originally a SOSD language that only later had a second diphthongization event happen to some stressed closed syllables.

    since Catalan and Occitan only diphthongize /ɔ/ before a palatal: Catalan avui, ull, nit (Old Cat. nuit), Occitan uèi, uèlh, nuèit. I don't know about Franco-Provençal or Northern Italian languages.

    French diphthongized closed syllable stressed ɛ in those same environments btw: tertiam > tierce, lectum > */lɛjto/ > /ljɛjt/ > lit, neptiam > nièce
     

    Swatters

    Senior Member
    French - Belgium, some Wallo-Picard
    possibly eastern Catalan (see below)

    What was he referring to? :confused:
    They propose to explain the reversal of /e/ and /ɛ/ in Eastern Catalan (piram > 'pɛrə, catēnam > kə'ðɛnə vs. herbam > 'erβə, tempus > tems) as the relic of a diphthongisation of /ɛ/ to /jɛ/, allowing /e/ to lower to /ɛ/, then /jɛ/ to monophthong to /e/:

    *nittam > *neta > *netə > *nɛtə > nɛtə (neta)
    *neptam > *nɛta > *njɛtə > *njɛtə > netə (néta)
     
    Last edited:

    Cenzontle

    Senior Member
    English, U.S.
    Additionally, your comments seem to clarify why "diptongacion" occurred mainly in nouns and verbs with certain initial consonants...mainly with bilabials. [#5 above]
    Probably best not to follow this detour.
    tuerto, duende, estiércol, suerte, chueco, cueva, güero, jueves,..
     

    Awwal12

    Senior Member
    Russian
    I think it may sound like logo (or something close) in rapid speech, but the underlying form is still /ˈlwego/. Diphthongization is as exactly as common in the colloquial as in the standard.

    Someone once drew attention to the fact that "el nuevo presidente" may also sound like "el novo presidente", and upon repeating it to myself various times it seems true. You can hear it in newscasts. Most likely a feature of rapid, natural speech. Hopefully someone will give a more precise answer.

    Also while aspirantization of coda s is quite common, I don't think it leads to complete deletion, it's still audibly different from ata.
    Basically if something has diphtongized it always can monophtongize back as well at some point. The General American /oʊ/ is pretty illustrative in that regard. While some may postulate that monophtongal realization of the phoneme resulting from the earlier /*o:/ and /*oʊ/ never disappeared in certain dialects to begin with, it clearly cannot be true for most dialects anyway. Moreover, in some words (like "know") the standard diphthong comes from the monopthong which resulted from monopthongization of basically the same dipthong (so we essentially have [oʊ]>[o:]>[oʊ]>[o]~[ö̞] here, with several intermediate stages and mergers along the way; all of that must have occured just during the last 5 centuries or so).
     

    Sobakus

    Senior Member
    Basically if something has diphtongized it always can monophtongize back as well at some point. The General American /oʊ/ is pretty illustrative in that regard. While some may postulate that monophtongal realization of the phoneme resulting from the earlier /*o:/ and /*oʊ/ never disappeared in certain dialects to begin with, it clearly cannot be true for most dialects anyway. Moreover, in some words (like "know") the standard diphthong comes from the monopthong which resulted from monopthongization of basically the same dipthong (so we essentially have [oʊ]>[o:]>[oʊ]>[o]~[ö̞] here, with several intermediate stages and mergers along the way; all of that must have occured just during the last 5 centuries or so).
    Did you read this somewhere? This is not how this works at all. Sound shifts have no memory and don't know what sound they come from. If something shifts back to a sound it originated from, one must assume that some new structural causes are responsible; for instance, if a variety of English transported overseas loses all of its diphthongs, no continuity with Middle English can be assumed and the cause must be, for example, a loss of the constraint on the long vowels' differentiated structure (the two parts may not be identical segments, as e.g. in Finnish). If all long vowels are shortened, it's a loss of bimoraicity.

    I don't see how Gen.Am. is illustrative in this respect, or why it clearly can't be true about the monophthongal realisation. The language simply never developed a new phoneme to fill the phonetic space left after the diphthongisation, probably because of the internal structure of English vowels (long vowels = diphthongs). This is the only reason they remain allophonic - just think of all the English vowels whose pre-GVS places have been occupied: you don't see "eye" monophthongising back to /i:/, or "how" to /hu:/. The pre-GVS allophones of these vowels have indeed disappeared (perhaps with some exceptions in the north of England), which makes it only more obvious that the source allophone of /o:/ has not.

    Moreover, language is a system of lects: a realisation can disappear from a dialect, but remain allophonic as long as that dialect remains part of the larger system where this realisation exists. Hearing these on the radio is enough.

    In the case of Spanish /ue/ vs /o/, it's a simple allegro speech synalepha (a hiatus elimination strategy) which is all over the place on word boundaries, but is generally restricted inside the word. It happens to be the same as the original sound as a function of the values of the two sounds involved (high back rounded X mid front), but I don't expect that the synalepha of /je/, if and when it happens, is going to be the /ɛ/ it originated from because the values of the two sounds (high front X mid front) are incompatible with an open front vowel as a synalepha outcome.
     
    Last edited:

    Awwal12

    Senior Member
    Russian
    Did you read this somewhere? This is not how this works at all. Sound shifts have no memory and don't know what sound they come from.
    This is precisely why they may get reverted. If A>B occured at some point, that fact alone has no power to prevent the subsequent B>A shift. And even though some shifts are more likely than other for various reasons (sometimes *ways* more likely), in the end it's a stochastic process, which cannot be precisely explained or predicted.
    The language simply never developed a new phoneme to fill the phonetic space left after the diphthongisation
    Which explains pretty much nothing, frankly. Yes, a shift which results in a phonemic merger is arguably less likely, but it is still an extremely commonb occurence. On the other hand, if the phonological system doesn't prevent some shift in any manner, it doesn't somehow mean that shift *will* happen.
     
    Last edited:

    Sobakus

    Senior Member
    This is precisely why they may get reverted. If A>B occured at some point, that fact alone has no power to prevent the subsequent B>A shift. And even though some shifts are more likely than other for various reasons (sometimes *ways* more likely), in the end it's a stochastic process, which cannot be precisely explained or predicted.
    What do you mean when you say that they get reverted? My point is that nothing gets reverted: a new sound shift occurs, and if its result coincides with the source, this does not imply any casual connection between the two. The synalepha of /ue/ in Spanish happens to coincide with its source; that of /je/ does not. In both cases no reversal is occurring: synalepha is not the reverse of diphthongisation in any structural sense as far as I can judge. The two processes may give the same result, or different results, depending on many factors. A true reversal would be a situation where exactly one thing switches back and then forth; but as far as I can see, this contradicts the basic principle of entropy. No matter how small a change, in a system as complex as human language you will never be able to revert exactly back to the state you started - even in the idiolect of one single person.

    Anyhow, do you have some American dialects in mind that lack the diphthongal /o:/, that you have reason to believe used to have it, and this can't be explained by external influence?
     
    Last edited:

    Awwal12

    Senior Member
    Russian
    My point is that nothing gets reverted
    You realize that's just the matter of how you put it, don't you? No kind of causation was ever implied above and I don't quite get how you have come to that idea. :confused:
    synalepha is not the reverse of diphthongisation in any structural sense
    There is no need to bring the structure up when we're simply considering a single phonetic shift.
     

    Sobakus

    Senior Member
    Ok, let me try to explain more clearly.
    Basically if something has diphtongized it always can monophtongize back as well at some point.
    Everything I wrote above was to illustrate why this sentence is wrong. It does establish a causation between one direction and the other: such causation does not exist. Not everything that has diphthongised can always monophthongise back, nor does it matter whether it had or had not diphthongised at all. As a byproduct of this, I find the statement
    The General American /oʊ/ is pretty illustrative in that regard.
    and everything that follows misleading and confusing. Not only do I think the thing you tried to illustrate is incorrect, I don't understand why you tried to illustrate it using the Gen.Am. /oʊ/, as I believe it illustrates precisely the opposite - the ability of the original allophone to linger on centuries after the actual phonemic restructuring. I'm sorry if I'm being forceful, but I truly am confused by what you're saying.
    There is no need to bring the structure up when we're simply considering a single phonetic shift.
    I don't believe it's possible to discuss phonetic shifts without considering phonology. Phonetic shifts are nothing but surface reflections of structural changes.
     

    Awwal12

    Senior Member
    Russian
    It does establish a causation between one direction and the other
    Er... how? It can be basically re-formulated as "if some phonetic shift is, in principle, possible, a reverse shift is also possible" (a shift to zero is an obvious exeption). It does NOT establish any connection between particular instances of the shifts.
     

    Sobakus

    Senior Member
    Er... how? It can be basically re-formulated as "if some phonetic shift is, in principle, possible, a reverse shift is also possible" (a shift to zero is an obvious exeption). It does NOT establish any connection between particular instances of the shifts.
    This precise formulation is what I think is false. Sound shifts progress in one direction because time and entropy progresses in one direction; they aren't reversible because they occur in highly complex, highly entropic systems. Let's take the Spanish example: an opening diphthongisation, whose structural reason was the elimination of mid-open vowels, has ran the course /ɛ > je/, producing two segments that already existed in the language: /j/ and /e/*. There is now no way to get an /ɛ/ out of these two segments. If I looked up some autosegmental descriptions and learned to write the process down formally in a pinch, you would see that for instance the feature [+open] of the [+front][+mid][+vowel] /ɛ/ simply cannot be derived from the combination of the features of /j/ and /e/, which both lack it; this feature has been lost and cannot be recovered anymore. If anything, the result of /j+e/ will be the addition of the feature [+close] to the vowel, giving /ẹ/. This is precisely what my opening statement means: sound shifts have no memory of which features they result in the loss of, and any shift that can result in the addition of this feature is coincidental to the sound shift that results in its loss. Language change is never zero-sum.
    * though /e/ itself results from the loss of the feature [+close]

    There may be some borderline cases where a shift in progress suddenly loses the conditions for its occurrence, and thus appears to have been reversed; but the same considerations as above tell me that there will always be some fall-out of this, some partial splits or mergers. Or it might conceivably become so substandard that the speakers consciously avoid it: I would be rather inclined to treat such examples as wholesale replacements of one grammar by another, and as such examples of external interference and not internal developments. It also needs to be very regular and transparent in this case. I would be interested to read a case study on this if anybody has one.
     
    Last edited:

    Awwal12

    Senior Member
    Russian
    Let's take the Spanish example: an opening diphthongisation, whose structural reason was the elimination of mid-open vowels, has ran the course /ɛ > je/, producing two segments that already existed in the language: /j/ and /e/*. There is now no way to get an /ɛ/ out of these two segments.
    Point taken. Still, with diphtongizations and monophtongizations the process in many cases goes in both directions nearly equally well.
     

    S.V.

    Senior Member
    Español, México
    It seems the issue was 'back' and 'reverse'. Awwal did say basically, but even in "one" direction, causal invariance (Stephen Wolfram) can still lead to the 'same' sorted string (i.e. a monophthong through a thousand different paths, if model runs for long enough)
    0414img82.png

    0414img81.png
    Then life gives you sayings like Everything new is a well-forgotten old, or a thousand words on Repetition and Difference.
     
    Last edited:

    Penyafort

    Senior Member
    Catalan (Catalonia), Spanish (Spain)
    By metaphonic diphthongization, they mean that it happened (at least in the examples they provide) in both open and closed stressed syllable, but that it was condition by vowel harmony: (Ischia Neapolitan) *por'kɛllu and *por'kɛlli > pur'tʃjeddə, while *por'kɛlla and *por'kɛlle > pur'tʃeddə.

    The rest of the chapter discusses what the link between those diphthongizations are, and whether the source is ultimately stressed syllable lengthening (which works well for the SOSD type, but less so for the others) or metaphony.

    They don't discuss it in the chapter, but at least for Walloon the distribution of diphthonged closed syllable ɛ and ɔ is more complex than what they present: in Western dialects like mine only ɛ diphtongize (bɛsta > bjɛs, fenestram > fɛrnjɛs, *perdutum > pjerdy, similare > *sɛmilare > *sjɛmler > ʃɛne), but ɔ doesn't (porta > pɔrt, scortea > (ɛ)skɔrʃ, costa > kɔs), while in central and eastern varieties, ɔ does (porta > (Namur) pwart, (Liège) pwɛt; scortea > N ʃwarʃ, L hwɛs; costa > N kwas, L kwɛs). This diphthongisation isn't uniform either: with a few rare exceptions, it only happens before sCV and rCV clusters, so bellam yield bɛl and not *bjɛl, contra becomes kõt and not kwɛ̃t, etc.

    Finally, the outcome of ɛ and ɔ in open syllable differ from those in closed syllables: ferum (proud) > fiːr while ferrum (iron, one of the rare exceptions I mentioned above) > fjeːr. Compare Spanish ferum > fjero; ferrum > fjero.

    Metaphonic diphthongization is also present in Iberia, specially in Aragonese and Asturian. Regarding Aragonese, and while traces of it can still be seen in certain words, toponyms are the best remaining attestation, specially in areas where the influence of Castilian didn't reach until the 20th century. Generally speaking, Ĕ-A and Ŏ-A would become ia-a and ua-a respectively (siarra < SĔRRA, lacuniacha < LACUNĔLLA, cuanga < CŎNCHA, cuasta < CŎSTA) while Ĕ-U and Ŏ-U mostly became ie-o and ue-o, as in Spanish.
     

    symposium

    Senior Member
    Italian - Italy

    Sobakus

    Senior Member
    Metaphonic diphthongization is also present in Iberia, specially in Aragonese and Asturian. Regarding Aragonese, and while traces of it can still be seen in certain words, toponyms are the best remaining attestation, specially in areas where the influence of Castilian didn't reach until the 20th century. Generally speaking, Ĕ-A and Ŏ-A would become ia-a and ua-a respectively (siarra < SĔRRA, lacuniacha < LACUNĔLLA, cuanga < CŎNCHA, cuasta < CŎSTA) while Ĕ-U and Ŏ-U mostly became ie-o and ue-o, as in Spanish.
    The examples you give are those of unconditional diphthongisation operative on any stressed mid open vowel, as in Castilian; in some varieties it's limited to open syllables. Metaphonic diphthongisation is conditioned by a (historically) following high vowel or glide u/i (typically in the ending), as in the Neapolitan examples.
     

    Sobakus

    Senior Member
    Point taken. Still, with diphtongizations and monophtongizations the process in many cases goes in both directions nearly equally well.
    I don't think this is true at all: again, these processes are non-zero sum and happen for structural reasons, not randomly. In the process their underlying representation changes so that features are added, deleted or combined in ways that no process exists that can add, remove or split them back. Perhaps if you familiarise yourself with autosegmental phonology and unerlying representations, this will become more apparent. I can compare this with lossy compression: every time you change something about a JPEG image and then save it, even if you "undo" it later, information is being lost. The Gen.Am. example illustrates creating copies of a single JPEG image and modifying them; later certain areas in the less-modified images can be cropped and pasted in order to fix the images that suffered the most compression. If the result looks the same, simply use the magnification tool. Nowhere in this is it possible to talk about genuine reversal after memory has been flushed, all because of the lossy compression.

    Likwise, language learning involves decompression of the signal and its encoding into the brain with lossy compression; its production also involves decompression (also lossy and with artifacts). It's impossible to arrive precisely at a previous state already due to the fact that you started with a derivative of that state. I think realising this is fundamental in understanding language change, and this is why I continue to protest against this confusion. It's not possible to change one single thing about a language and flip the switch back to reverse it in the next generation - it's always a game of Chinese whispers. More importantly, under normal circumstances a structure like language has consistent systemic pressures that force it to evolve in one direction, and flipped switches stay flipped. If you see an outcome that resembles a previous state, this is no reason to suspect that the underlying structure of the system has stayed the same. This is what I mean when I say what looks like reversals are actually coincidental. I think the example with the two Spanish diphthongs is enough illustration for this. If anything needs to be added, it's that even the synalepha of /ue/ does not result in the original [+open] vowel /ɔ/, but in the normal Spanish [-open] vowel /o/; there surely are other structural elements such a monophthongisation does not recreate, for example length, because even allophonic vowel length no longer exists in Spanish.
     
    Last edited:

    Awwal12

    Senior Member
    Russian
    I don't think this is true at all: again, these processes are non-zero sum and happen for structural reasons, not randomly. In the process their underlying representation changes so that features are added, deleted or combined in ways that no process exists that can add, remove or split them back.
    Phonetic shifts are fundamentally, well, phonetic, not phonological. Phonology is just what can make them more or less likely. While phenomena like the Great Vowel Shift or parallel developments in related languages demonstrate that structural factors may greatly influence the subsequent phonetic shifts, the shifts still remain stochastic by nature.
     

    Sobakus

    Senior Member
    Phonetic shifts are fundamentally, well, phonetic, not phonological. Phonology is just what can make them more or less likely. While phenomena like the Great Vowel Shift or parallel developments in related languages demonstrate that structural factors may greatly influence the subsequent phonetic shifts, the shifts still remain stochastic by nature.
    The only way I can interpret it is that you're equating phonetic shifts to regularization of articulatory mistakes. I've never seen any linguist doubt the phonological, structural basis of sound change, and I find it unthinkable. An interaction of physiological and phonological factors? Absolutely, but never what you're saying. Do you have any references for such a view?
     

    Awwal12

    Senior Member
    Russian
    I've never seen any linguist doubt the phonological, structural basis of sound change
    Except that would essentially make phonetic shifts 100% predictable and the resulting linguistic divergence pretty much impossible, which clearly contradicts the facts we can observe.
     

    Sobakus

    Senior Member
    Except that would essentially make phonetic shifts 100% predictable and the resulting linguistic divergence pretty much impossible, which clearly contradicts the facts we can observe.
    Not any more than life being based on evolution and DNA replication makes life on Earth 100% predictable. Come on, this is not serious. You aren't the kind of person that doesn't understand the difference between the basis of something, and the overarching web of secondary factors that influence it to a greater or lesser degree. I take it I won't see any linguist saying what you're saying today, though.
     

    Awwal12

    Senior Member
    Russian
    Not any more than life being based on evolution and DNA replication makes life on Earth 100% predictable.
    A phonological structure is a structure. Neither evolution nor DNA replication are, they're stochastic processes, as much as biological life itself. I don't see where the analogy is supposed to be seen here.
     

    Sobakus

    Senior Member
    A phonological structure is a structure. Neither evolution nor DNA replication are, they're stochastic processes, as much as biological life itself. I don't see where the analogy is supposed to be seen here.
    So you believe that phonological structure is some hard, immutable, unchanging structure, without any uncertainty or dynamic (processual) elements, and therefore can underlie no stochastic process? To dispel such a notion it's enough to remind you that a linguistic structure exists as a continuously modified process in an indiviual's brain, and is replicated like DNA by transmission (language learning), during which an individual tries to match their own brain process - by simple guessing, which we call baby babbling (although underlied by some universal language faculty) - to what they hear in such a way that what they produce back is interpreted by the listener (the parent) in the way they intend; that this transmission is never perfect hardly needs mentioning. When a baby learns a language, they guess what switches are on and off in their parent's heads, and some of these switches can and do get changed later.

    This is an individual - a linguistic system is an immense hive of such individuals, who all have their switches in slightly different combinations. A system like this is a structure and a process at the same time, and the more complex it gets, the more zeroes you can prefix to the 0.1 probability that you can predict its behaviour - this is reasearched by system theory.
     
    Last edited:

    Awwal12

    Senior Member
    Russian
    So you believe that phonological structure is some immutable, unchanging structure
    It's obviously changing - as a *result* of phonetic shifts (ultimately, of fluctuations in articulatory reproduction). At any given moment, however, it can be described in a sufficiently accurate and complete fashion. The statement "phonological structure is what makes phonetic shifts happen" makes as much sense as "road signs make the cars move".
     

    Sobakus

    Senior Member
    It's obviously changing - as a *result* of phonetic shifts (ultimately, of fluctuations in articulatory reproduction).
    Yet you have no proof or citations for this. You think this for the same reasons uncivilised peoples believe that the stochastic probability of rain increases as a result of their dancing for it, or why we believed desease to be stochastically caused by "bad air" 150 years ago. This is nothing new - when you have no means to describe something as a system, it takes but a simple lack of imagination to claim that there's no system and it's all random or magic. I'm disappointed that views such as these are held on an etymology forum.
    At any given moment, however, it can be described in a sufficiently accurate and complete fashion.
    I don't think you've even begun to explore theoretical linguistics if you think anything has been described in such a way. We aren't even close to that.
    The statement "phonological structure is what makes phonetic shifts happen" makes as much sense as "road signs make the cars move".
    Phonological structure is the road along which cars move. Cars don't move where there's no roads. A car might be found there once in a while, but that car doesn't exist to a traffic minister. It doesn't form a system. You still know that cars don't drive in the middle if the sea like you know that the phoneme /oʊ/ will not be monophthongised into an /i/. If you build a road, there's a 100% probability that cars will be found there, as long as there are cars around. Building a road makes cars appear on it by no magic at all.
     

    Awwal12

    Senior Member
    Russian
    This is nothing new - when you have no means to describe something as a system, it takes but a simple lack of imagination to claim that there's no system and it's all random or magic.
    Oh, but there is a system in phonetic shifts (otherwise typology of such shifts wouldn't even exist). It just has not that much to do with the phonological system of the particular language.
    You still know that cars don't drive in the middle if the sea like you know that the phoneme /oʊ/ will not be monophthongised into an /i/.
    Which precisely has nothing to do with phonology, since the extremely low likehood of that shift comes simply from the articulatory perspective and doesn't depend on the particular language at all.
    Yet you have no proof or citations for this.
    If one and the same phonological structure may experience entirely different developments, it necessarily means that this structure isn't the only factor that defines the results at the very least. Calling a structure a fundamental driving force behind any changes (its own changes included) already sounds absurd enough, I hope.
     

    Sobakus

    Senior Member
    Oh, but there is a system in phonetic shifts (otherwise typology of such shifts wouldn't even exist). It just has not that much to do with the phonological system of the particular language.
    I have asked that you support this statement with references on numerous occasions. Do you understand the concept of supporting your words with references? Or do you instead believe in repeating the same assertion of your own invention until the other party rolls their eyes and gives up trying to reason with you?
    Which precisely has nothing to do with phonology, since the extremely low likehood of that shift comes simply from the articulatory perspective and doesn't depend on the particular language at all.
    And the low likelyhood of rain this month comes from the fact that we haven't danced hard enough for it and has nothing to do with meteorology. Am I doing this right?
    If one and the same phonological structure may experience entirely different developments, it necessarily means that this structure isn't the only factor that defines the results at the very least. Calling a structure a fundamental driving force behind any changes (its own changes included) already sounds absurd enough, I hope.
    Have some decency - it was you who stated that "phonetic shifts are fundamentally, well, phonetic, not phonological". I replied with "An interaction of physiological and phonological factors? Absolutely". The conclusion of this short paper, which I don't think you've even bothered opening, reads:

    The current study eliminates one of the theoretically possible answers to this question. Speakers implement duration increases in fricative portions in word-final position, so the constraint on fricative lengthening cannot be an articulatory one, but must be phonological in nature.
    Not only have you done nothing to substantiate any of your fanciful assertions despite being repeatedly asked to; you're trying to turn the tables on me. How little respect - forget for me, for yourself! - do you have to hold a discussion in this way?

    Way to go, way to go. In retrospect, the statement you're repeating reveals such poor understanding of phonological theory that I should have simply dropped a reference like this and be done with it instead of naively trying to bridge that unbridgeable gap: Kiparsky P. 2003, The phonological basis of sound change

    At least since Jakobson 1929, evidence has been accumulating that sound change itself, even the exceptionless kind, is structure-dependent in an essential way.
    [...]
    I argue that vowel shifts are another type of natural sound change whose explanation, on closer inspection, depends on the structural status of the triggering feature in the system, specifically on whether the feature is specified in the language’s phonological representations or active only at the phonetic level.
     
    Last edited:

    Penyafort

    Senior Member
    Catalan (Catalonia), Spanish (Spain)
    The examples you give are those of unconditional diphthongisation operative on any stressed mid open vowel, as in Castilian; in some varieties it's limited to open syllables. Metaphonic diphthongisation is conditioned by a (historically) following high vowel or glide u/i (typically in the ending), as in the Neapolitan examples.

    The examples I gave were conditioned by the final vowel: ia for -A, ie/io for -U.

    It also happened with monophthongs.

    CUMBA > coma (Catalan coma, French combe)​
    CUMBU > cumo (Catalan cóm)​
    FURCA > forca​
    FURCU > furco, fulco (Catalan forc)​
    LUMBA > loma​
    LUMBU > lumo (Catalan llom, Spanish lomo, Portuguese lombo)​
    PETROSA > petrosa​
    PETROSU > petruso (Catalan pedrós, Spanish pedroso)​
     

    Swatters

    Senior Member
    French - Belgium, some Wallo-Picard
    Right, but what the quote was discussing were the triggers for diphthongization, not the shape of the resulting diphthong. From what you describe of Aragonese, metaphony affects the result of the diphthongisation, but not whether it happens or not (i.e. (theoretical examples) vetulum gives *viello, while vetulam gives *vialla) In southern Italy, we see a diphtong before historical high vowels (vetulum > viecchio), but a monophthong before historical low vowels (vetulam > vecchia). The processes are similar, but still show a significant difference (in Maiden's categorisation, Aragonese would be a member of the second category (systematic diphthongisation) while Neapolitan would be a member of the third (metaphony conditioned diphthongisation)
     

    Penyafort

    Senior Member
    Catalan (Catalonia), Spanish (Spain)
    Ok, I see now. I don't find the terminology chosen to be the best for this, but now I got it.
     

    Sobakus

    Senior Member
    The examples I gave were conditioned by the final vowel: ia for -A, ie/io for -U.

    It also happened with monophthongs.
    I'd actually never heard of a Romance language that had metaphony on top of diphthongisation, so I went looking around and found Ferguson T. (1976) A History of the Romance Vowel Systems Through Paradigmatic Reconstruction casually lying under a pile of socks on LibGen and ended up reading about the Rumanian vowel system. The book is probably not that up-to-date, being basically structuralist, but it's an exceptionally clear and systematic treatment of the developments.

    Anyhow, Rumanian apparently has a similar situation, only with open-syllable diphthongisation preceding metaphony in open and closed syllables. I haven't found any mention of such a situation for Aragonese, however - all the descriptions I've seen simply mention that diphthongisation has varying outcomes, like Ŏ > /uə ue ua uo/, with variable accentuation, often arguing that such a state is an indication of its archaic nature. Could you point me to a source with further examples and treatment? Non-English is fine as long as I can get my eyes on it on the Web. I've found a study on Asturian, which also seems to have both diphthongisation and a separate metaphony, but it excludes diphthongs from the investigation.

    From what I'm reading now in Maiden/Smith/Ledgeway (2011) The Cambridge History... p.134 and Ledgeway/Maiden (2016) The Oxford Guide... p.656, Loporcaro disagrees with this explanation of Rumanian, but he's got his own Open Syllable Lengthening agenda. If this also exists in Aragonese (and Asturian-Leonese), this improves Ferguson's position.
     
    Last edited:
    Top