Xephyr wrote:WeepingElf wrote:The point is, of course, that such an assumption is, well, an assumption. The h-index notation has the advantage that it doesn't suggest a particular set of phonetic values which may turn out to be wrong.
Yes, that's the usual defense, and for me personally it has never seemed a good one. Lots of things in historical linguistic reconstructions are assumption...
lots. Like the nature of phonation distinctions in stop series. Or the nature of two contrasting coronal fricatives. If you are someone who is looking at a proto-language to begin with it's generally assumed that you know that. But we tend not to give protolanguages fricative inventories of s1 s2 or stop inventories of p1 p2 t1 t2 k1 k2... so why not just make an orthographic simplification with PIE laryngeals?
Actually, you do see that kind of notation. Proto-Kartvelian is typically reconstructed with three series of central affricates and sibilants. The voiceless sibilants, for example, are commonly written
*s s₁ š (somewhat illogically). Subscript or superscript numerals are also quite common in comparative afro-asiatic linguistics. Since there is quite a bit of controversy in semitic linguistics over how the original pronunciation of the coronal fricatives/affricates, it's not uncommon to write
*s₁ s₂ s₃ for traditional
*š ś s. And then there's Germanic
*ē₂ (and
*ē₁), to give another example.
I don't think it's a good idea to "just assume a pronunciation" if the pronunciation is a point of controversy or if it's generally uncertain. You could argue that it's just an orthographic convention, and is by no means meant to signify that there is agreement on how the phoneme was pronounced, but in practice it will probably be impossible to make everyone agree upon a convention if it's biased towards one side of a debate (it's hard enough even with a neutral proposal). Someone who strongly believes that
*h₁ was [ʔ],
*h₂ [ħ] and
*h₃ [ʕ] is likely to not accept
h,
x and
xʷ. The effect is, of course, multiple competing conventions which is generally a bad thing. Note that some (but not all) supporters of the various glottalic theories use different notation for the PIE plosives, so in hindsight, using dʱ etc. was probably not the best idea. It's more valuable that the conventions are widely agreed upon than that they are aesthetically pleasing.
It's better to use symbols that don't strongly suggest a certain pronunciation. This doesn't necessarily mean using subscript numerals, however. I personally dislike this convention too, both because it's ugly and because it can be hard to remember which symbol is which. But it's possible to use "neutral" diacritics or use letters that are not strongly associated with a certain pronunciation. The semitic convention of writing emphatic consonants with a dot (ṭ, ḳ, ṣ etc) is a good example of this. There is disagreement over whether the emphatics were originally ejective, or pharyngealized, or something else, but since the dot doesn't really suggest any particular pronunciation, different authors can still agree on how to write them.
Writing
*h₃ as
xʷ is a particularly bad idea since it quite unambiguously specifies a labialized pronunciation, something which is by no means certain and which some linguists would probably argue against. Apart from the unlikelyhood that the convention will be widely adopted, the specificity in itself is actually misleading as it might suggest to the reader that labialization is widely agreed upon. Using
x for either laryngeal is not a bad idea, however. Even if there is an IPA-value associated with the letter, it is widely used for many different pronunciations in the world's languages (and x is often used to signify something unknown).
Obviously, there are a number of neutral ways in which the laryngeals could have been written, many of which I would personally prefer over the subscript numerals. But it's unlikely that the current convention will change...
h₁ h₂ h₃ (h₄) H
x ẋ ẍ (ẍ̇) X
ḥ ẖ ḩ (ȟ) h͓
q x c (ȝ) ḥ
ɿ ƨ з (ч) x
