Talk:Germanic spirant law

[Untitled]

Something this article definitely needs is an explanation of clusters ending in PGmc -d- rather than -t-. This was affected as well, as shown by Gothic: nasidēdun from nasjan, but wissēdun (< *wit-dēdun, from probably a PIE cluster d-dh) from witan. It is in fact highly doubtful that the original consonant of the preterite was -t- especially considering its relationship to 'do', and it's more likely that the change to t was a result of assimilation to a preceding voiceless consonant. However even this does not explain Gothic baúhta, baúhtēdun (bought, < *bug-dēdun) from bugjan, nor Old English meahton (might, < *mag-dēdun) from magan (with the change gd > ht). These would have had two voiced consonants (originating from PIE gh-dh, two voiced aspirates), making devoicing through assimilation alone unlikely. I can't see an exact solution here, but the article really has to account for -d- if it's going to be any use at all. --CodeCat (talk) 13:45, 21 November 2008 (UTC)

The article as it stands is deliberately basic and limited to things which are not controversial. It sounds to me like you are exploring details which are still in debate. If you wish to add a section on on-going research at the bottom, please do, but do give good references, because this shouldn't become OR. But whatever you do, please make sure it keeps a form which a layperson can easily understand (with a little effort to read into the subject), which is the aim at the moment. --Doric Loon (talk) 18:50, 21 November 2008 (UTC)

Well the problem is, I don't have any references. I was kind of wondering about the stuff I mentioned as well, and I was hoping someone else could add something which would also be helpful for me. That's why I posted this. --CodeCat (talk) 12:34, 22 November 2008 (UTC)

Morphological conditioning

The change affecting dental consonants is generally assumed to have been a separate phenomenon, and was already a part of Proto-Indo-European phonetics, since other Indo-European languages show similar results. It seems to have only occurred in cases where the second /t/ was the first part of a suffix; geminated /tt/ that occurred within a single morpheme remained. Evidence from Germanic as well as other Indo-European languages such as Latin confirms this. For example, Latin edere "to eat" shows the past participle esus "eaten" from earlier *ed-tus. But a geminate /tt/ is preserved in both Gothic and Latin atta "father".

The assumption that sound laws are sometimes inherently sensitive to morphology is problematic, and therefore such explanations should be avoided; it is better to assume that sound laws are "dumb" and occur as generally as possible (in fact, it is quite likely that sound laws initially even ignore word boundaries), and that the seeming sensitivity of sound laws to morphological environments is due to later analogical restoration.

In the specific case, the assumption that morpheme-internal /tt/ developped differently from /tt/ which arose out of morpheme contact is even unnecessary: It is generally agreed that PIE did not even have geminates (compare 2nd singular present *h₁esi of the copula instead of *h₁es-si, despite a morphological boundary), and therefore there cannot have been a lexeme *atta in the first place (which is an implausible reconstruction for other reasons, as well). Clearly, the more economic assumption is that whenever two dentals met, the same development involving a sibilant happened, and that the similar-looking words for "father", "old man" etc. in several IE languages are simply independent innovations from baby-talk or loanwords (the Slavic etymon may well be a loan from Gothic). Father-words following the pattern pa(-pa), ba(-ba), ta(-ta), da(-da) and mother-words after the pattern ma(-ma) or na(-na) (the second occurrence of the consonant may be geminated, and the first occurrence or the final vowel, or even both, omitted) are dime a dozen and extremely common in the languages of the world, and therefore it is foolish to uncritically assume them to be old and inherited from PIE. --Florian Blaschke (talk) 08:44, 2 September 2011 (UTC)

Assimilation

Can this law actually be called an instance of assimilation? As it changes combinations of stops into combinations of fricative + stop, it should actually be called a type of dissimilation, shouldn't it? --Florian Blaschke (talk) 09:09, 2 September 2011 (UTC)

I suppose it is dissimilation but there is more to it. Grimm's law would normally turn plosives into fricatives, but in a series of two obstruents the second never becomes a fricative, only the first one (unless it's /s/ which is already a fricative). So historically it's a special case of Grimm's law, not a separate instance of dissimilation. CodeCat (talk) 10:47, 2 September 2011 (UTC)

Are you sure that /st/ is part of the spirant law? How about /ks/, is it part of the spirant law? We need more sources. I've found something: This handbook of Old High German considers Primärberührung to be the voice assimilation within consonant clusters which occurs prior to Grimm, i. e., all cases of /D(^h)T/ or /D(^h)s/ (strictly speaking, however, only the cases /bt/, /gt/, /dt/ and /gs/ are listed, because Indo-Iranian displays Bartholomae's Law in the case /D^hT/ and apparently also /D^hs/, hence the voice assimilation cannot be of PIE age in this case), as well as "a second variant", namely /TD(^h)/ and /sD(^h)/, while this handbook on Middle High German evidently even considers clusters consisting of /D(^h)D(^h)/ and even nasal group + /D(^h)/ part of the phenomenon. However, original /pt/, /tt/, /kt/, /st/ and /ks/ are emphatically not part of the phenomenon according to the sources, as no voice (nor any other kind of) assimilation has taken place (pre-Grimm). So this article is getting it totally wrong. Also, strictly speaking, the German term should be primärer Berührungseffekt or Primärberührungseffekt. --Florian Blaschke (talk) 15:18, 2 September 2011 (UTC)

The solution in this case would be to split this article into two articles, namely one on the Germanic spirant law (whose scope still remains to be determined using sources), and one on the Primärberührungseffekt. --Florian Blaschke (talk) 15:51, 2 September 2011 (UTC)

OK, thinking it over, I think you are right about the spirant law. The development of /ks/ > /xs/ is completely in tune with Grimm, hence cannot be part of the law if we define it as "exceptions from Grimm involving clusters with spirants". However, /st/ would seem to be part of it. Whether /tt/ is part of it is unclear to me. As far as I know, it is controversial how /tt/ developped in Germanic, and what the Pre-Proto-Germanic (i. e., pre-Grimm) reflex was. If it was /ss/ all along, nothing would have happened at all, and Grimm would never even have been relevant for the group. --Florian Blaschke (talk) 16:01, 2 September 2011 (UTC)