Weak rules (was: swallow vs. nighingale, Common Romance)

From: Grzegorz Jagodzinski
Message: 50902
Date: 2007-12-15

---- Original Message ----
From: alexandru_mg3
To: cybalist@yahoogroups.com
Sent: Sunday, October 28, 2007 11:53 AM
Subject: Re: [tied] swallow vs. nighingale, Common Romance

--- In cybalist@yahoogroups.com, "Grzegorz Jagodzinski"
<grzegorj2000@...> wrote:

>> In fact, we are NOT saying about rules at all here. I have already
>> said that
>> the change may be called at most "a weak rule" or a tendency, as
>> only some
>> words developped due to it.

> Please do not use such theories as "weak rules" this remember me
> other of your formulations like "anyhow related"...

Why should I stop using them? Because YOU do not like them?

I do not intend to stop using formulations which describe facts
adequately. I would like to remind you that the conception of weak
rules is not my personal invention. Namely, not all linguists are
blind confessors of Neogrammarian theories of only exceptionless
rules (called "laws" in addition). "Such theories as <weak rules>"
are not pseudoscience then, they are not my personal obsession either
(I use them only when needed, and I do not deny existence of
traditional rules which do not allow exceptions), so there is no
reason not to use them. And in addition, this is not a mailing list
on do's and don't's, or on should or shouldn't's - but on
Indo-European languages. So, it would be better to discuss on such or
another view than on what such or another user should or shouldn't
do.

Weak rules are phonetic changes that comprise only some instances of
words - while the others remain unchanged (in the same conditions).
In other words: not individual ("idiosyncratic") and not fully
regular changes. You believe that there are not weak rules - your
problem. I can see enough examples and I believe that they exist.
Contrary to you, I do not tell you "please use this" and "please do
not use that". Such pleases are below the level of normal
discussion and I do not like them very much.

You may also believe that, let's say, the Spanish nn > ñ and ll > [lj]
phonetic rules are without any relation with one another (and, that in
general, phonetic processes do not influence one another). I will
term the two processes "anyhow related" - because such is my choice.

I will not discuss with someone's beliefs - as I believe (sic!) that
it would be out of sense. Instead, I will make a webpage (when the
time allows) with examples of irregularities which may be explained
as results of weak rules - and then we will be able to return to this
subject.

As for now, only excerpt:

1. Various dissimilations in many lngs which are only rarely fully
regular (more often r ... r > l ... r or r ... l, but also e.g.
Spanish madrileño < madrideño).

2. The third palatalization in Slavic. I presented my view in this
point some time ago as well as my doubts on the proposed phonetic
limitations of the process; now I can add one more doubt, namely
Slavic zaje,cI "hare" need not come from *zajink- as it was
suggested, compare Finnish jänis "hare", the stem jäniks-, which may
be a borrowing from Slavic and which may prove *za-jenk- instead of
*zajink-.

There also exists a plausible explanation of this dark word. Namely,
it is probably related with Slavic words for "spider", with no single
reconstructible protoform (Polish paja,k suggests *pa-eNkU while Czech
pavouk, Russ. pauk suggest *pa-oNkU). The stem *eNkU / *oNkU meant "to
curve" (cf. the curved spider's legs, or the curved silhouette of a
sitting hare), see Lith. anka "a loop, a knot".

3. Posttonic nonfinal front vowel (e, i) > a in Spanish, like in
pájaro (passerem), cuévano (cophinum), pámpano (pampinum), muérdago
(mordicum); according to rules, this vowel should disappear rather.

4. A similar weak rule for initial not stressed vowel, like in añade
< enade < *inaddit, balanza < bilancem, trabajo < *tripa:lium.
Naturally, we may speak of assimilation (to the next "a") here - but
it is an irregular phonetic process (or: a weak rule: it sometimes
occurs, sometimes does not) as well.

5. A weak rule *au > a (instead of regular o) in Spanish, like in
agosto < augustum (dissimilation?) or in Old Spanish ascuchar Btw.,
irregular vowel changes occurred twice in modern escuchar < Old
Spanish ascuchar < auscultat; note that this time none of the changes
were assimilations, dissimilations or changes due to frequence.

6. Prothetic initial v- (ex nihilo) before o- in Slavic (while regular
before y- < u:-), ex. *vonja "a smell" (in Polish: won') < IE h2anH-.

7. In Slavic: *pt > t, even if in other words *pt > st (the same
hesitation for former *bt). It does not seem to be dialect-dependent.
The development *pt > t is thought to be south Slavic but it is not
true: stryjI < *ptru:- "father's brother" in all dialects, cf. Latin
patruus. Also in Old Polish both niec' and nies'c' existed side by
side, < *nept- (the exact meaning is disputed, one of possibilities
is "grandson of uncle / aunt"). More examples to come.

8. In Slavic, *s > x sometimes (weak rule!) in other positions when
after *r, *u, *k, *i (plus after *l like in pIlxU "Glis, grey
squirrel" < *pl.sos, which is not accepted by all scholars).
Examples: paxati < *po:s- "to blow, to smell", koxati < *kos- "to
love" (cf. kosnoNti "to touch"). Some try to explain such a "x" as a
result of "expressive" development of *s. But what does it mean
"expressive" indeed? Is it something really different from a weak
rule? I do not think so.

See also *xmura ~ *smura "cloud" (cf. Polish pochmurny and Russian
pasmurnyj "cloudy" and also Old Russian xmuryj "dusky"), and possibly
xoditi "to go, to walk (frequently or without aim)" - cf. Greek hódos
< *sod-.

9. Winter's rule (aka Winter's law) in Slavic. We can observe
exceptions in both directions, i.e. vowel lengthening when not before
a plain voiced ("glottalized") stop (like in *se^kti "to cut, to
hack" - *sekyra "axe" with short *e may not belong here, it be an
Italic borrowing) and without such a lengthening before such a
consonant, like in voda "water" < IE *wodo:r (**vada expected), or
possibly also in xoditi (see above).

10. Psilosis and related phenomena (h-metathese) in Classic Greek. It
appears in some words only, like in *a:uso:s > *a:uho:s ~ *ha:uo:s >
e:os (Homeric) ~ heo:s (Classic) "dawn".

11. In Greek w- > h- was a weak rule (while w- > 0- in other words),
ex. hénnymi, hestía < *wes- (cf. Cret. we:ma, Latin vestis etc.),
hésperos (cf. Latin vesper, Greek dialectal wespérios), historía (<
*wid-tor-). Even if we assumed *w ... s > *h ... s regular, see also
Attic hédna "marriage gifts" : Hom. éedna (< *(e)wed-, h- in Attic is
irregular - cf. Slavic ve^no "marriage portion" < *wednom, with
regular lengthening due to Winter's rule). For w- > 0- in Greek, cf.
épos, étos, eméo:, eíkosi (while dial. wi:kosi), élpis, eréo:, éar (<
*wesar with *w ... s!), oînos, oîkos etc.

12. Yet other weak rules in Greek (Attic and Ionic dialects) are in
the development of sonants r., l. - namely, sometimes they yield ra,
la, while ar, al in other instances. For example: amaldýno:, bdállo:,
árktos, tharsýs, kardía (Attic) - but bladarós, tharsýs, kradíe:
(Ionic), kráteros (Attic and Ionic).

13. A similar problem with the m. sonant + a laryngeal is seen
ká:mno: ~ kékme:ka (a:m ~ me: < *m.H) or in da:mázo: ~ dma:tós, dmô:s
(a:ma ~ ma: < *m.H, with no clear rules).

14. *wl.- > *lu-, possibly as early as in Common IE, cf. Latin lupus,
Greek lykos "wolf" (while without such a change in plenty of other
words).

15. *kW > *p in some words in various dialects (cf. words for "4",
"wolf", and perhaps for "river": *akWa: ~ *apa:). Note that Latin has
*kW > p in lupus "wolf" even if *p > kW can be observed in other
words, like quercus, quinque or coquere (regular (?) distant
assimilation of *p to *kW).

16. The famous s-mobile. The presence of forms with s- and without s-,
sometimes in the same dialect, is irregular by itself. In addition,
*(s)t- and *(s)k- seem to come from different pre-IE sounds (from
Nostratic affricates) than *t-, *k-. It it was true, the alternation
*st- ~ *t- (like in *taur- ~ *staur- "bull, aurochs") and *sk- ~ *k-
would be irregular also fonetically (in other words, ex. Nostr. *c^ >
IE *st > *t would be a weak rule).

17. Prothetic vowels in Greek. Sometimes they are really traces of
laryngeals - but sometimes they show ad hoc, without clear rule.
Sometimes we have both forms, with the prothesis and without it. No
Neogrammarian "strict" rule can describe such instances at all. While
prothetic vowels always develop before r- - it is a classical,
Neogrammarian rule -, the simplest solution (so, the scientific one)
is that sometimes (a weak rule!) they also develop before initial l-,
m-, n-. Examples: elakheîa ~ lakheîa "small", aleípho: "to anoint" ~
liparós "anointed", amaldýno: "to make soft" ~ malakós "soft",
possible also ané:r, ánthro:pos ~ dró:ps "man".

18. The prothesis before *w- could result IE *Hw-. However, the
prothesis somethimes disappers without clear rules. See eérgo: (<
*ewergo:) ~ (w)ergázomai "to work", eíkosi ~ wi:kosi "20", eédna ~
hédna (< *(e)wedna).

19. Other prothetic vowels also appear and disappear without strict
rules:
ekhthés ~ khthés "yesterday",
ethélo: ~ thélo: "I want",
ekeî ~ keî "there",
aspaíro: ~ spaíro: "I breathe hard",
Obriareús ~ Briareús (btw., without r ... r dissimilation),
iktís ~ ktís "marten", here probably also iktînos "kite (bird)"
(and "a kind of wolf"?),
ikhthy^s "fish" with prothesis - but khthô:n "land" without
prothesis (cf. also khamaí "on the ground" with irregular reduction
of th, due to frequence, so not Neogrammarian as well).

20. IE *sk- preserves in Slavic, as a rule (as s^c^- before a front
vowel and as sk- elsewhere). But sometimes it yields *ks- (> x-).
There even exist some dublets, like *skUrbU ~ *xUrbU "treasure" (the
second form only in Polish personal and place names Charb and
related), perhaps *skUrpa ~ *xUrpa (Polish skarpa "escarp", if not
borrowed ~ Old Polish charpa "unevenness"), *skorU "fast, prone,
prompt" ~ *xUrtU "greyhound", also *xoNprU (< *sk-, in Old Polish
che,pr of 1564) ~ *kuprU "rump, bottom, coccyx".

21. The variation *oN ~ *u in Slavic. I does not mean the late
process *oN > u in most Slavic languages except Slovene (*oN > o) and
Polish - but the original alternation. The simplest solution is that
once existed a weak rule which transformed some *oN's into *u's
(example: l/e,k ~ l/uk, both Polish and both with the meaning "arch",
among others; more examples to come).

22. Reduction processes like aphairesis, synkope, apokope,
hyphairesis in Greek (also haplology / hapaksepia - all known also in
other languages) which are all irregular (i.e. they sometimes take
place, sometimes do not). Examples: skorakízo: "to drive away (to
ravens)" < *eis-korakizo:, tórnos < tóronos "compasses", Berníke: ~
Bereníke:, Doric da:miergós ~ da:mo:rgos (both in the same dialect,
the first with hypheresis, the second with regular contraction) <
*da:miowergós (Homeric de:mioergós), tetrakhmon (in inscriptions) ~
tetrádrakhmon "a four drachmas coin", *amphiphoreus > amphoreús
"amphora". Not all such examples can be explained with the help of
frequency (which is, in addition, not aknowledged to be a valid
reason of phonetic changes by radical Neogrammarians).

22. -d- > -r- in Spanish, like in sur "south" (more examples - ? or
it is just "exception"?).

23. -ss- > -s's'- (-ssi-) in only several words and in only some part
of the Romance language territory.

24. The relation between IE *enk-, *lenk-, *leng-, *deng- (all with
the meaning "to curve", see 2) is another example (all the stems can
be found in Slavic). Other examples of similar variations are:
*nHg^u- ~ *dnHg^u- ~ *lnHg^u- ~ *gnHg^u- "tongue",
*ak^ru- ~ *dak^ru- ~ *lak^ru- "a tear".
Whatever we would say on these examples, the
correspondences of the variations are similar but irregular. Even if
we supposed a stop-lateral *dl- in IE in these words, the development
of this sound is not regular (for example, both lingua (old dingua)
and gingiva in Latin).

See also German Träne < trahan "tear", with tr- < *dr- < *dl- ?
Yet another example with *dl- (> *dr-, *l-, *0-) may be Greek dráko:n
"dragon", Latin lacerta and probable also locusta, Lith. skerys "locust",
Slavic as^c^erU "dragon". These words show even more irregularities - but
Romance forms are also irregular, cf. Italian lucertola (a > u ?), Spanish
and Portuguese lagarta or Spanish langosta.



> Gregorz, can you understand that Latin ss > Spanish s and that we
> need to find the origin of ssi in pájaro elsewhere?

I can understand that you cannot imagine processes that are not fully
regular. Maybe the examples cited above will allow you to understand
that such processes DO EXIST. Be patient and wait for more examples.

Grzegorz J.



___________________________________________________________
Now you can scan emails quickly with a reading pane. Get the new Yahoo! Mail. http://uk.docs.yahoo.com/nowyoucan.html