User:Jackwolfroven/Reconstruction hypotheses

This page contains phonological, grammatical, etc. features of extinct/ancient languages which are not generally accepted in the linguistic community but which I believe were factual, including evidence to support them.

Indo-European edit

Italic vocalism edit

Italic/Latin <ă> /a/ usually reflects Proto-Indo-European *e.

Evidence edit

"Plain voiced" row was glottalized/ejective edit

The PIE "plain voiced" stops (*b, *d, , *g, *gʷ) were phonologically ejective or glottalized consonants. The "aspirated" row (*bʰ, *dʰ, *ǵʰ, *gʰ, *gʷʰ), then, contained phonologically unaspirated and voiced stops. Id est:

Proto-Indo-European consonant segments
Labial Coronal Dorsal Laryngeal
velar post-velar labiovelar
Nasal *m *n
Plosive voiceless *p *t *ḱ [k] *k [q] *kʷ  
ejective (*b [pʼ]) *d [tʼ] [kʼ] *g [qʼ] *gʷ [kʷʼ]  
voiced *bʰ [b] *dʰ [d] *ǵʰ [ɡ] *gʰ [ɢ] *gʷʰ [gʷ]  
Fricative *s *h₁, *h₂, *h₃
Liquid *r, *l
Semivowel *y [j] *w

Evidence edit

  • PIE voiceless stops tend to become "plain voiced" when placed at the end of a word:

In many modern languages, including English, aspirated or voiceless stops tend to become glottalized (or replaced completely with a glottal stop) word-finally, preserving any aspiration only in careful speech. Danish stops undergo a similar process.

  • Winter's law and Lachmann's law, lengthening vowels before the "plain voiced" stops in closed syllables in Proto-Balto-Slavic and in Latin, respectively.

The lengthening of vowels suggests that the vowels were "interrupted" and then lengthened for compensation. This interruption could have been caused by glottalization. Note that the change did not involve the voiceless or "aspirated" stops. Also note that Lachmann's law has been disputed because of its operation on the deep phonological structure. This could alternatively have been acting on a still-present glottalized stop, whose adjacency to a voiceless stop caused it to become unvoiced when the glottalized consonants underwent glottal slack to become full voiced elsewhere.

  • The reflexes of the "aspirated" stops:
PIE ǵʰ gʷʰ
Anatolian ? ? ? ? ?
Tocharian p t k k
Hellenic kʷʰ
Kharpert Armenian p t ? k ?
Sebastia Armenian ? ?
Indo-Iranian ĵʰ gʰ, ǰʰ gʰ, ǰʰ
Albanian b d, dh /ð/ d, dh /ð/ g, gj /ɟ/ g, z
Balto-Slavic b d ź g g
Germanic b ~ β d ~ ð g ~ ɣ g ~ ɣ b ~ w
Celtic b d g g
Italic f ~ β f ~ ð h ~ ɣ h ~ ɣ f ~ ɣw

The only branches to show an aspirated reflex are Hellenic, Indo-Iranian, and Armenian, which were likely either in geographical proximity or in a Sprachbund together (it is even proposed that they share a genealogical connection later than PIE). The rest show an unaspirated voiced articulation, with the exception of Tocharian. The Italic branch additionally shows fricative reflexes, and these reflexes appear very similar to unrelated sound changes in more modern languages (e.g. Attic Greek /b/ to Koine Greek /β/ to Byzantine Greek /v/; Latin /b/, /d/ to Castilian Spanish /β/, /ð/; Latin /b/ to Italian /v/). It is typologically improbable that all of these branches lost the aspirated series in all environments very early in their respective historical sound changes.

Tocharian edit

In Pre-Tocharian or Proto-Tocharian, there was at least a five-way contrast between plosives when the main palatalization occurred: *p, *t, *tʼ (which would be *d according to the traditional view), *k, and *kʷ, since PIE *d palatalized differently in Tocharian from how *t and *dʰ did. There may have been more phonemic contrasts at the time of the palatalization sound change, but there must have been at least these five. Preceding a front vowel, the following behavior can be observed in the stops:

Pre-palatalization *p *t *tʼ/d k
Palatalization *p' (pʲ) *c ś ś

This could possibly be an indication that PIE *t and *dʰ merged before *d also did, or that the full "voiced aspirate" series merged onto the voiceless series before the "plain voiced" also did, although this is not necessarily true. What it does indicate is that the "voiced aspirate" dental stop palatalized in the same way as the voiceless dental stop, but that the "plain voiced" dental stop was distinguished from both through palatalization. This suggests that the "voiced aspirate" series was phonetically more similar to the voiceless series than to the "plain voiced" series.

Armenian VOT edit

The Armenian reflexes of the plosives are more complicated than can be summarized simply by the above table. The different reflexes among the dialects are shown below, taking the dental plosives as an example:

Correspondence in initial position
Indo-European *d *dʰ *t
Sebastia d
Erevan t
Istanbul d d
Kharpert, Middle Armenian d t
Malatya, SWA d
Classical Armenian, Agulis, SEA t d
Van, Artsakh t t

The consonants transcribed <dʱ> are breathy voiced.

Here is my hypothesis concerning the development of these assorted dialectal reflexes:

Development in initial position
Inherited series Armenian aspiration of voiceless stops Dialectal loss of voice contrast Greco-Aryan aspiration isogloss reaches Armenosphere Dialectalizations Modern dialects
*d /tʼ/, *dʰ /d/, *t
*tʼ, *d, *tʰ
*tʼ, *d, *tʰ
*d, *dʱ, *tʰ
*d, *dʱ, *tʰ
*d, *dʱ, *tʰ
*d, *d, *tʰ
*t, *dʱ, *tʰ
*t, *dʱ, *tʰ
*t, *d, *tʰ
*tʼ, *t, *tʰ
*d, *t, *tʰ
*d, *tʰ, *tʰ
*t, *t, *tʰ
Middle Armenian, Kharpert
*d, *t, *tʰ

This model suggests the existence of an isogloss which explains the aspiration seen in Hellenic, Indo-Iranian, and Armenian. Such an isogloss would have originated in roughly the same location, but at a later time than, the Satem isogloss, since this isogloss did not reach the Balto-Slavic branch as Satemization had. In almost all cases, the aspiration which occurred allowed for a chain shift to occur, causing the ejective/glottalized plosives to develop modal voice.