Guild of Linguists

Click here to edit subtitle

Forums

Post Reply
Forum Home > The Language Room > Proclitics and Reduction

KathTheDragon
Site Owner
Posts: 381

It's well-established that the D'ni prepositions could stand as independent words or be cliticised onto the noun they govern. When they are proclitic, it is also well-established that they can be reduced, with their e being weakened to ', or completely elided altogether. The definite article re- also sometimes undergoes this reduction, as does the conjunction ga in a few instances. However, to my knowledge, the rules regarding this reduction have never been fully determined, since there seem to be examples and counter-examples to almost every formulation. Yet, I think I've come up with an explanation which explains the whole dataset, assuming I have correctly identified all the errors.


 

The most important observation is that ga, a word not normally known for being reduced, is reduced in a small number of texts - Atrus' Prayer, the page from Gehn's Journal, and the navaot Gateroom text. In Atrus' prayer, there are four instances of ga, three of which are prefixed, and two of those are reduced. It should be noted that the unreduced attestation stands before h, which is known for blocking reduction in another word where reduction is expected: behapo, an infinitive. Further, the unprefixed attestation is conjoining clauses, a position where prefixation is never attested. In Gehn's Journal, there is only one instance of ga, which is prefixed and reduced. In the Gateroom text, there is only one instance of ga, which is prefixed and reduced. These texts all have in common that they were composed post-Fall, making them the latest strata of native D'ni texts. This implies a chronological progression: gah- > g'- (g- before re-) which is perhaps analogical to the prepositions, as discussed below. There is additionally the attestation of g'rerovtí in the DLG, but this may be an error, considering the attestation of garekor in FMR.


The reduction of re- appears in the Gateroom texts in two instances: r'rinæltav (note that I consider r'inæltav to be a mistake for this word) and r'ráwé. All other instances of re- are unreduced. This pattern can be predicted by a single exceptionless rule: re- > r'- before r. This rule also generally holds in Gehn's Journal, where there is a single exception: r'aça. However, further down the page we find rearíutav and treaça. On the basis of these, I consider r'aça to be a mistake for reaça. There are two instances of reduction in information provided by Cyan/Rawa: r'erem and r'éc'dé. It may be the case that the form of D'ni on which Cyan based their standard had the rule re- > r'- before vowel, but whether this is paralleled elsewhere is uncertain. The Kenen Gor text has a fully reduced r- in xrelaþtí, but non-reduced re- in xre'íst. We also have fully reduced r- in the names Relto and Renyaloth. It's possible that standard D'ni had reduction only before e and é, which would make the presence of reendétan in Gehn's Journal analogical. There is a single attestation of unreduced re- before r in rerúé in Aitrus' Map, which is either a mistake or an indication that the rule re- > r'- before r is not standard D'ni.


The situation with the prepositions is interesting. All the prepositions attested before re- are fully reduced, except for ne- and one instance of me-, and would probably also be so before erþ-, though this is only attested for te-. In the case of be-, as a proclitic, it is always attested reduced except in behapo, as noted above. This is also the only case of prefixation to a word beginning in h, which raises the possibility that h blocks reduction, as in the attestation of gahúcéþom in Atrus' Prayer. There are three attestations of be- as a separate word, all of which belong to Gehn - two in his journal, the other uttered by him in his schoolroom speech. With te-, we have numerous instances of a reduced proclitic, and a separate word. There is also a single attestation of an unreduced proclitic in teflin, which should probably then be regarded as a mistake. se- is only attested once as a separate word. xe- is frequently attested as an unreduced proclitic. fe- is attested once as a reduced proclitic in f'túgo in Aitrus' Map, and twice as an unreduced proclitic in fesev in Gehn's Journal and femagentíom in Atrus' Prayer. me- is attested twice as an unreduced proclitic in med'ní and mepræd, and many times as a separate word. ne- is not attested (except for the one instance before re-).


The above discussion of the prepositions is hard to turn into rules for reduction, though it seems that for these, reduction is generally performed lexically. One clear rule is that all prepositions are reduced before re-, and possibly more generally before r. At least te- is reduced before erþ-, though this could be extended to all prepositions and before e generally. Furthermore, reduction seems to be blocked before h. Otherwise, be- and te- are always reduced, xe- and me- are not reduced, and the status of the other prepositions is unknown.

--

Moula KI: 00005310
DI KI: 00205116
deviantART: kathaveara
tumblr: kaththedragon

Grand Master of the Guild of Linguists


October 28, 2015 at 12:27 PM Flag Quote & Reply

Talashar
Member
Posts: 30

I think it's more likely that D'ni allows a certain amount of variation than that all the unexpected values are mistakes.


de and le also have reduced forms, though there aren't many examples of either.  The alternation of de with d' in devokan versus d'nee and d'mahlah is probably lexical (though for all we know there's an invisible difference in stress between mahlah and vokan) and doesn't cast much light on the other reduction processes,  The fact that reduction occurs just before nasals is almost certainly due to the miniscule sample size!


The two instances where the Perfect prefix le is reduced are both before vowels: l'a(h)rta(h)em in Atrus's prayer and probably l'ayl[...] inside the cover of the Book of Atrus.  (But le'emee is used in Kadish's note.)

--

Talashar Geltahn; Ki 183867 An overview of D'ni grammar | My books

October 28, 2015 at 8:39 PM Flag Quote & Reply

KathTheDragon
Site Owner
Posts: 381

I wasn't sure about including those two, since there's just not enough data, and it's all conflicting. I only included the lesser-attested prepositions because of the parallels with the other prepositions.


I strongly doubt that the exceptions to the rules I've discovered are simply variation. I only adduced three definite errors, one likely error, and one word that can be explained with recourse to dialectal differences. And in each case, I cited parallels that make a good case for them being errors (and in the case of r'rinæltav vs. r'inæltav, I think it's a closed case that one of these is a mistake, and my choice of r'inæltav being incorrect is made with recourse to other attested forms). And to be honest, 4 or 5 mistakes is a pretty good rate, considering just how many attestations there are! If these were simply variation, why are there so few deviant forms? I'd rather expect the variants to be vastly more common.

--

Moula KI: 00005310
DI KI: 00205116
deviantART: kathaveara
tumblr: kaththedragon

Grand Master of the Guild of Linguists


October 29, 2015 at 8:13 AM Flag Quote & Reply

You must login to post.

Oops! This site has expired.

If you are the site owner, please renew your premium subscription or contact support.