Unrecoverable segments in Proto-Nuristani reconstructions

While at first glance reconstructing Proto-Nuristani seems to be straightforward, it is important to exercise caution in reconstructing unrecoverable segments that are not found in the modern Nuristani reflexes.

(For abbreviations and sources, see this post.)

An example is given below:

PIIr. *ȷ́ʰarH-ai̯na- "golden > red" (Halfmann 2025) + *-ka- > Kt. (W) zëřë́, NE zëňë́, (SE) j̈ëňë́

As mentioned in the source above, a PNur. *j̈arenáka- could be reconstructed. The pretonic correspondence of Kt. /ë/ in all dialects suggests earlier *e. But there are no traces of the final syllable, nor the sequence *-ar-, so these segments have to be inferred from other Indo-Iranian languages (such as OIA hári-).

It is still possible to deduce the final vowel of some Proto-Nuristani etyma, particularly through forms descending from an extended *-ka-.

Consider the cognate set for "milk":

PNur. *j̈ara- "milk" (Halfmann 2025)

> Ash. (M/T) <zọ̄>, (S) <z′ō>

> Kt. (W/NE) zu, (SE) j̈u

(PNur. *j̈araka- + diminutive -g) > Pr. ḍug

> NKal. (Wg/Zh) <zōr>, (N) <zor>

Here, all forms except Prasun come from *j̈ára-, with regular outcomes of stressed *a in a monosyllable (Ash. /o/, Kt. /u/, NKal. /o/). The Prasun form instead comes from *j̈aráka-, where the change of *a > /u/ is regular, revealing that the Proto-Nuristani etymon was an a-stem (one could also posit a PNur. o-stem due to the regular change of *o > Pr. /u/, but this is unsupported within Indo-Iranian.)

But not all forms can be explained by an extension with *-ka-. Other segments can be posited, such as intervocalic *y, *w, or any other consonant subject to lenition. Even different final vowels can be inferred, and they can produce the attested reflexes, such as the cognate set for "heart":

PIIr. *ȷ́ʰŕ̥dai̯a- "heart"

> Ash. (M/T) <ži'dī>, (S) <žid′i>

> Kt. (W/NE) zërë́, (SE) j̈arë́

> Pr. zər ~ zir

> NKal. (Wg) <zō>, (Zh) zo, (N) zö

> Tr. (G) <ź′o>

We can see some differing correspondence sets here, though all of them show regular reflexes of *r̥d: Ashkun, Prasun /i/ and Nuristani Kalasha, Tregami /o/ here point to stressed *r̥, but Katë (W/NE) /ë/ ~ (SE) /a/ clearly implies pretonic *r̥ merging with *a. This implies that the word was originally stressed on the second (penultimate) syllable in Proto-Nuristani, i.e. *j̈r̥dáya-, from which the Katë form would regularly derive. The other forms could theoretically come from *j̈ŕ̥de-, with regular *aya > *e (cf. *trai̯as "three" > Kt. tre) while still obeying penultimate stress, or even a secondarily thematized form *j̈ŕ̥d-a- < *ȷ́ʰŕ̥d-a-. It is even possible to conclude that the Kt. forms come from *-ka-extended *j̈r̥dáka-, since intervocalic *k and *y normally disappear. We can therefore only posit the existence of *y or final vowels via inference from other Indo-Iranian forms, and judging from the wider Indo-Iranian cognates, a reconstruction of *j̈r̥dáya- is the most likely.

This issue is the most problematic for forms with apparently no non-Nuristani cognates, such as the cognate set for "leek" below:

PNur. *pāndā- "leek, onion"

> Ash. (S) <p′â˜t>

> Kt. (W) pon ~ pun (f.), (SE) pon (f.)

> NKal. (Wg/Zh) <pān>, (N) <pân> ~ <pān>

It is clear from the stressed vowel correspondences (Ash. /a/, Kt. /o/, NKal. /ā/) that the first vowel is *ā. We can also infer the presence of *nd via its Ashkun cognate (normally > /n/), but here final *d devoices to /t/. The only segment that cannot be recovered is final *ā, which is inferred from the feminine gender of the Kt. forms, and there are no known *k-extensions to determine the final vowel. The PNur. reconstruction given here is a minimalist reconstruction with the fewest unreconstructable segments. By inference, it is perfectly possible to posit other reconstructions, such as *parāndā-, *patāndā-, or any other segment containing *a and a consonant subject to lenition.

Thus, it is important to consider all possible segments when reconstructing Proto-Nuristani, particularly if a word has no external cognates.

References

  • Halfmann, Jakob (2025). The Diversification of Indo-Iranian and the Position of the Nuristani Languages. Wiesbaden: Reichert.

 

Comments

Post a Comment

Popular posts from this blog

Is Proto-Nuristani *ã justifiable on the basis of the Comparative Method?

One phoneme mentioned in Halfmann (2025) is the presence of * ã in Proto-Nuristani reconstruction. The relevant mention of this claim is found in page 94: With this in mind, the word for ‘wool’ can be reconstructed as either PNur. *warnaka-/*warrãka- or *wərnaka-/*wərrãka-, but certainly not as *ūrnaka- as would be required for a parallel with the Indo-Aryan development. Here, we will try to justify the existence of a distinct Proto-Nuristani vowel phoneme * ã by looking at the cognate sets below, and determine if any of the Proto-Nuristani reconstructions above agree with the application of the Comparative Method: Sources For practical reasons, only attested forms alongside a Proto-Indo-Iranian reconstruction will be cited below. Richard Strand's website (https://nuristan.info/lngFrameL.html) will be used alongside sources for Ashkun (Morgenstierne 1929, 1934), Katë (Halfmann 2024), Prasun (Buddruss & Degener 2017), Nuristani Kalasha (Degener 1998, Tāza 2017), and Tregami...
Read more

Preliminaries on Nuristani comparative reconstruction: Single consonant correspondences

The Nuristani languages are a group of Indo-Iranian languages spoken in the Hindu Kush mountains in Afghanistan and small parts of Pakistan. Compared to Indo-Aryan and Iranian, Nuristani is first attested only in the 19th century, and has absorbed a large amount of Indo-Aryan loanwords since the Middle Indo-Aryan period. The Nuristani languages are Ashkun, Katë, Prasun, Nuristani Kalasha, and Tregami. The position of the Dameli language is unknown within Nuristani or Indo-Aryan; it has a large number of Nuristani-derived vocabulary. While there is a serious and urgent need for further research, especially with the groundbreaking works of Georg Morgenstierne and Jakob Halfmann, it is best to start with the principles of the Comparative Method, a powerful tool for analyzing language relationships. The comparative method  The comparative method analyzes the sound correspondences of two (or more) related languages or dialects, and uses that to establish hypothetical earlier forms to es...
Read more

Preliminaries on Nuristani comparative reconstruction: Consonant clusters

In this post, we explore the historical development of consonant clusters in the Nuristani languages. The last two posts covered single consonant and vowel correspondences. However, similar to vowel developments, consonant clusters have not received much attention before Halfmann (2025)'s recent book on the position of the Nuristani languages. This post will attempt to find regular reflexes of Proto-Indo-Iranian consonant clusters in the Nuristani languages. Some of these were already discussed by Halfmann (2025). Due to lack of information, not all Proto-Indo-Iranian consonant clusters are represented here, though that may change with further research. Sound correspondences We will start with normative (excluding conditioned reflexes) consonant cluster correspondences for the Nuristani languages, based on clearly inherited forms. For practical reasons, only attested forms alongside a Proto-Indo-Iranian reconstruction will be cited below. Richard Strand's website (https:...
Read more