On Sat, Apr 27, 2013 at 07:38:22AM -0700, Roger Mills wrote:
> --- On Fri, 4/26/13, H. S. Teoh <hsteoh@...> wrote:
> I'm trying to settle on a consistent glossing scheme for Tatari Faran
> interlinears. In TF, certain sequences of sounds are considered
> "unpleasant", and will mutate in order to preserve euphony (in the
> ears of native speakers, anyway, not necessarily corresponding to a
> foreigner's sense of aesthetics). This usually happens across word
> boundaries, for example:
> huna + na -> hunan da
> asusu + sei -> asusei
> Does this happen in every such case? I.e. ...VnV + na > ...Vnda, or
> > ....ssei > ...sei??
In the former case, yes:
VnV + na > Vn da
However, Vn + na is unchanged:
tun + na > tun na
So the process seems quite specific to -VnV.
In the second case, the rule is actually:
sVsV + sei > sVsei
Though whether it's a _rule_ is debatable, since currently there's only
one word that fits the rule, _asusu_. Nevertheless, there is a similar
rule with the subordinative verbal suffix -as:
arap + -as > arapas
akaisu + -as > akaisatas
In this case, one may formulate the rule sV + -as > satas.
> > Then it's a case of final-V deletion and -CC- sandhi, IMO, but I'm
> > not sure how Leipzig handles that. Maybe:
> huna + na
> 2Pl + RCPMASC
> hun + na
The rule doesn't apply in the second case, actually. _hun + na_ would
still just be _hun na_. Only when a vowel follows the final _n_ does the
rule take effect. It's also specific to the case clitics _na_, _nei_,
and _no_; so you have:
huu na hena naritai muin
1SG RCP:MASC and cheer_up FIN
And I cheered up.
*huu na henan daritai muin.
OTOH, you have:
karen hunan da tsuni tinka aba ira.
karen huna na tsuni tinka aba ira.
shoe 2PL RCP:MASC find conifer under FIN
Your(pl) shoes were found under the conifer.
> This is not dissimilar to sandhi phenomena in Kash, e.g.
> karun + mi
> lord 1SgPoss
> karumbi 'my lord' (i.e. like Engl. "m'lord")
Nice! It sounds similar to the kind of euphony rules TF has. :)
> rum + fasan
> CAUS hot
> to heat s.t.
> Or (more like your exs since there's deletion involved:)
> yurun + nahan
> PLACE EAT
> yur + nahan
> (in ththe first case rum- behaves exceptionally, which I won't go into
> here. The point is that you can't have *rundraka because of a
> phonological constraint against 2 /r/s in successive syllables.)
Interesting! I like the sound of the sandhis. Maybe I should make more
euphony rules in TF. :) Besides the VnV + na > Vn da rule, there's also
the contraction of _ei_ into a short vowel when it occurs in two
adjacent monosyllabic words:
sei + ei > si'ei
mei + sei > misei
mei + sei + ei > mei si'ei
In the third case, _mei_ doesn't contract because the contraction
between _sei_ and _ei_ already eliminated the long vowel _ei_ from being
adjacent to _mei_, so there no longer a need to contract.
> añ + raka
> NOML big
> bigness, size
> Final nasals in compounds and derivs. always assimilate to the
> following C --with exceptions-- e.g.
> karun + ni
> Lord 3sPOSS
> his lord(ship)
> this is because regular *karundi would be identical to karundi <
> /karun + ti/ (2s Poss).
Interesting. I've also tried to avoid the most egregious cases of
homophony in TF, but decided that it's permissible when it happens
between a genitive form and a case particle, for example:
huna + na > hunan da
karen huna-n + na > karen hunan da
shoe 2PL-GEN RCP:MASC
In the first case, the surface form _hunan da_ is analysed as _huna +
na_, whereas in the second case, the isomorph _hunan da_ is analysed as
_hunan + na_.
(Incidentally, this means that the rule is actually more complex than
I've stated earlier; it's actually both of:
VnV + na > Vn da
nVn + na > nVn da
Vowel deletion happens in the first case, not the second, but in both
cases _na_ shifts to _da_.)
Not all such cases are ambiguous, though; with words like _pasanan_,
another rule kicks in when forming the genitive:
pasanan + -an > pasanaran
So the surface forms of "town + na" is quite distinct from "town-GEN +
pasanan + na > pasanan da
san pasanaran + na > san pasanaran na
san pasanan-an + na > san pasanaran na
person town-GEN RCP:MASC
> How does one gloss such phonological mutations? The first case is
> especially troublesome, because _hunan_ on its own is a genitive of
> the pronoun _huna_, but in this case, it's merely a disguised form of
> its unmodified stem. So, should it be glossed as:
> hunan da
> huna na
> 2PL RCP.MASC
> or as:
> hunan da
> huna- n_da
> 2PL RCP.MASC
> According to the Leipzig glossing conventions, the second line should be
> a morpheme-by-morpheme breakdown; how should the base morphemes _huna_
> and _na_ be indicated, and how should they align with the orthographic
> So apparently the orthographic text should come last. Personally, I
> think it should come first since that's the form the reader is trying
> to analyze. Mais que sais-je?
Hmm. According to:
under Rule 2, 2nd para, it states that one may add another line at the
beginning containing the unmodified text. In my understanding, the last
line should be the free (English) translation? So something like this:
hunan da tsuni jibin ipasanaran ka ira
huna na tsuni jibin i-pasanan-an ka ira
2PL RCP:MASC find child PART-town-PART ORG:MASC FIN
You(pl) were found by the city boy.
(The circumfix i-...-an is an appositive, so _jibin ipasanaran_ is "city
boy" as opposed to _jibin pasanaran (GEN)_ "boy of (a particular)
Or did I misunderstand the rules?
Give me some fresh salted fish, please.