The multitude of
Middle Iranian languages and peoples indicate that great linguistic diversity must have existed among the ancient speakers of Iranian languages. Of that variety of languages/dialects,
direct evidence of only two has survived. These are: •
Avestan, the two languages/dialects of the
Avesta (the
liturgical texts of
Zoroastrianism). •
Old Persian, the native language of a southwestern
Iranian people known as
Persians.
Indirectly attested Old Iranian languages are discussed
below. Old Persian was an Old Iranian dialect as it was spoken in southwestern Iran (the modern-day province of
Fars) by the inhabitants of
Parsa, Persia, or
Persis who also gave their name to their region and language. Genuine Old Persian is best attested in one of the three languages of the
Behistun inscription, composed , and which is the last inscription (and only inscription of significant length) in which Old Persian is still grammatically correct. Later inscriptions are comparatively brief, and typically simply copies of words and phrases from earlier ones, often with grammatical errors, which suggests that by the 4th century BCE the transition from Old Persian to Middle Persian was already far advanced, but efforts were still being made to retain an "old" quality for official proclamations. The other directly attested Old Iranian dialects are the two forms of
Avestan, which take their name from their use in the
Avesta, the
liturgical texts of indigenous Iranian religion that now goes by the name of
Zoroastrianism but in the Avesta itself is simply known as
vohu daena (later:
behdin). The language of the Avesta is subdivided into two dialects, conventionally known as "Old (or 'Gathic') Avestan", and "Younger Avestan". These terms, which date to the 19th century, are slightly misleading since 'Younger Avestan' is not only much younger than 'Old Avestan', but also from a different geographic region. The Old Avestan dialect is very archaic, and at roughly the same stage of development as
Rigvedic Sanskrit. On the other hand, Younger Avestan is at about the same linguistic stage as Old Persian, but by virtue of its use as a sacred language retained its "old" characteristics long after the Old Iranian languages had yielded to their Middle Iranian stage. Unlike Old Persian, which has Middle Persian as its known successor, Avestan has no clearly identifiable Middle Iranian stage (the effect of Middle Iranian is indistinguishable from effects due to other causes). In addition to Old Persian and Avestan, which are the only
directly attested Old Iranian languages, all Middle Iranian languages must have had a predecessor "Old Iranian" form of that language, and thus can all be said to have had an (at least hypothetical) "Old" form. Such hypothetical Old Iranian languages include
Old Parthian. Additionally, the existence of unattested languages can sometimes be inferred from the impact they had on neighbouring languages. Such transfer is known to have occurred for Old Persian, which has (what is called) a "
Median" substrate in some of its vocabulary. Also, foreign references to languages can also provide a hint to the existence of otherwise unattested languages, for example through toponyms/ethnonyms or in the recording of vocabulary, as
Herodotus did for what he called "
Scythian" and in one instance,
Median (
σπάκα "dog").
Isoglosses Conventionally, Iranian languages are grouped into "western" and "eastern" branches. These terms have little meaning with respect to Old Avestan as that stage of the language may predate the settling of the Iranian peoples into western and eastern groups. The geographic terms also have little meaning when applied to Younger Avestan since it is not known where that dialect (or dialects) was spoken either. Certain is only that Avestan (all forms) and Old Persian are distinct, and since Old Persian is "western", and Avestan was not Old Persian, Avestan acquired a default assignment to "eastern". Further confusing the issue is the introduction of a western Iranian substrate in later Avestan compositions and redactions undertaken at the centers of imperial power in western Iran (either in the south-west in Persia, or in the north-west in Nisa/Parthia and Ecbatana/Media). Two of the earliest dialectal divisions among Iranian indeed happen to not follow the later division into Western and Eastern blocks. These concern the fate of the Proto-Indo-Iranian first-series
palatal consonants, *ć and *dź: • Avestan and most other Iranian languages have
deaffricated and depalatalized these consonants, and have *ć >
s, *dź >
z. • Old Persian, however, has fronted these consonants further: *ć >
θ, *dź > *ð >
d. As a common intermediate stage, it is possible to reconstruct depalatalized affricates: *c, *dz. (This coincides with the state of affairs in the neighboring
Nuristani languages.) A further complication however concerns the consonant clusters *ćw and *dźw: • Avestan and most other Iranian languages have shifted these clusters to
sp,
zb. • In Old Persian, these clusters yield
s,
z, with loss of the glide *w, but
without further fronting. • The
Saka language, attested in the Middle Iranian period, and its modern relative
Wakhi fail to fit into either group: in these, palatalization remains, and similar glide loss as in Old Persian occurs: *ćw >
š, *dźw >
ž. A division of Iranian languages in at least three groups during the Old Iranian period is thus implied: • Persid (Old Persian and its descendants) • Sakan (
Saka,
Wakhi, and their Old Iranian ancestor) • Central Iranian (all other Iranian languages) It is possible that other distinct dialect groups were already in existence during this period. Good candidates are the hypothetical ancestor languages of
Alanian/Scytho-Sarmatian subgroup of Scythian in the far northwest; and the hypothetical "Old Parthian" (the Old Iranian ancestor of Parthian) in the near northwest, where original *dw > *b (paralleling the development of *ćw). ==Middle Iranian==