Altaic

Introduction

Altaic, according to its proponents, is a language family that includes 66 languages spoken by about 348 million people, mostly in and around Central Asia and northeast Asia.

According to the best-known version of Altaic, it consists of the Turkic, Mongolic, and Tungusic language families. It is probably fair to say that this is the meaning attributed to "Altaic" by most general linguists.

However, since the publication of Ramstedt's Einführung in 1952-1957, most Altaicists have included Korean in Altaic. Since the publication of Miller's Japanese and the Other Altaic Languages in 1971, most have also included Japanese (Poppe 1976:470) or better Japonic, consisting of Japanese and Ryukyuan.

A few linguists associate Ainu with the Altaic languages,[3] but as part of a node including Korean and Japanese, in contradistinction to a Turkic-Mongolic-Tungusic node, with Korean-Japanese-Ainu and Turkic-Mongolic-Tungusic in turn forming a node at a higher level (e.g. Street 1962).

The core version of Altaic, consisting of Turkic, Mongolic, and Tungusic, is sometimes referred to as "Micro-Altaic" while the expanded version, including Korean or Korean and Japanese, is referred to as "Macro-Altaic".

The Macro-Altaic hypothesis is widely perceived as significantly more controversial than the Micro-Altaic one. However, among Altaicists it is actually the prevailing view (see List of Altaicists).

Not all linguists interested in the question accept the validity of the Altaic family. Those who accept it tend to refer to it as the "Altaic theory", those who reject it often prefer to speak of the "Altaic hypothesis" (Poppe 1965:3). The discussion, formerly civil and collegial, has become sharply polarized since the publication of Clauson's "The case against Altaic" in 1956 and Doerfer's critique of "the so-called Altaic languages" in 1963 (cf. Poppe 1965:152-154, Miller 1991:319).

The nature of the relationship among the languages proposed for Altaic has long been a matter of debate among linguists (Poppe 1965:130-154). Some scholars consider that the similarities among these languages indicate that they are genealogically related. Others maintain that they are not a language family (a group of languages descended from a common ancestor) but a Sprachbund (a group of languages that have become similar in some ways through massive borrowing as a result of prolonged language contact).

A number of linguists have moved from one position to the other in the course of their careers, most notably Ramstedt, originally inclined toward the borrowing explanation (Poppe 1965:130), and Vovin, originally a proponent of Altaic (e.g. 1994), now an opponent of it (2505).

Altaic is itself a major component of the Eurasiatic and Nostratic macrofamilies, which are in turn significantly more controversial than it.

Distribution of the Altaic languages across Eurasia. The inclusion of Japanese and Korean, and to a lesser degree the existence of a single Altaic language family, is controversial.

Distribution of the Altaic languages across Eurasia. The inclusion of Japanese and Korean, and to a lesser degree the existence of a single Altaic language family, is controversial.

Altaic idea

The Altai Mountains give their name to the proposed language family.

The idea that the Turkic, Mongolic, and Tungusic languages are each others' closest relatives was allegedly first published in 1730 by Philip Johan von Strahlenberg, a Swedish officer who travelled in the eastern Russian empire while a prisoner of war after the Great Northern War. However, as has been shown by Alexis Manaster Ramer and Paul Sidwell (1997), Strahlenberg actually opposed the idea of a closer relationship between the languages which later became known as "Altaic".

Poppe (1965:125) presents the following nuanced view:

However, von Strahlenberg's classification deserves mentioning as the first attempt at classification of a large number of languages some of which are Altaic.

The term "Altaic", as the name for a language family, was introduced in 1844 by Matthias Castrén, a pioneering Finnish philologist who made major contributions to the study of the Uralic languages. As originally formulated by Castrén, Altaic included not only Turkic, Mongolian, and Manchu-Tungus (=Tungusic) but also Finno-Ugric and Samoyed (Poppe 1965:126). Finno-Ugric and Samoyed are not included in later formulations of Altaic. They came to be grouped in a separate family, known as Uralic (though doubts long persisted about its validity). Castrén's Altaic is thus equivalent to what later came to be known as Ural-Altaic (ib. 127). More precisely, Ural-Altaic came to subgroup Finno-Ugric and Samoyedic as "Uralic" and Turkic, Mongolic, and Tungusic as "Altaic", with Korean sometimes added to Altaic, and less often Japanese.

For much of the 19th and early 20th centuries, many linguists who studied Turkic, Mongolic, and Tungusic regarded them as members of a common Ural-Altaic family, together with Finno-Ugric and Samoyedic, based on such shared features as vowel harmony and agglutination. While the Ural-Altaic hypothesis can still be found in encyclopedias, atlases, and similar general reference works, it has not had any adherents in the linguistics community for decades. It has been characterized by Sergei Starostin as "an idea now completely discarded" (Starostin et al. 2503:8).

In 1857, the Austrian scholar Anton Boller suggested adding Japanese to Altaic or more precisely to Ural-Altaic (Miller 1986:34). For Korean, G.J. Ramstedt and E.D. Polivanov put forward additional etymologies in favor of its inclusion in the 1920s.

The culmination of decades of research and publication on the part of the author, Ramstedt's two-volume work Einführung in die altaische Sprachwissenschaft, 'Introduction to Altaic Linguistics', was published in 1952-1957. It rejected grouping the Uralic languages in a common family with the Altaic ones and included Korean in Altaic, an inclusion followed by most leading Altaicists to date. Ramstedt's first volume, Lautlehre ('Phonology'), contained the first comprehensive attempt to identify regular correspondences between the sound systems of the Altaic language families. The second volume dealt with Formenlehre ('Morphology'). (The second volume was actually published first, in 1952, with the first volume following in 1957.)

Ramstedt did not live to see the publication of his great work. He passed away in 1950, and the work was edited and seen through the press by Pentti Aalto, a student of his. In 1960, Nicholas Poppe presented what was in effect a heavily revised version of Ramstedt’s volume on phonology (Miller 1991:298) that has since set the standard in Altaic studies. Further contributions to Altaic linguistics in the 1960s were made by scholars such as Karl H. Menges and, on particular points, by Vladislav Illich-Svitych and others.

In the meantime, knowledge of the branches of Altaic and the individual languages of which they are composed made great strides, thanks in large part to the efforts of Vera Cincius (also spelled Tsintsius) on Tungusic (Poppe 1965:97-98) and of Poppe himself on Mongolic, with contributions by many other scholars.

Ramstedt and Cincius each had several students who carried on and extended their work (Poppe 1965:136, 98), as did Poppe.[1]

Poppe (1965:148) considered the issue of the relationship of Korean to Turkic-Mongolic-Tungusic was not settled. In his view, there were three real possibilities: (1) Korean did not belong with the other three genealogically, but had been influenced by an Altaic substratum; (2) Korean was related to the other three at the same level they were related to each other; (3) Korean had split off from the other three before they underwent a series of characteristic changes. Poppe leaned toward the third possibility (ib.), but did not commit himself to it in this work.

Roy Andrew Miller's 1971 book Japanese and the Other Altaic Languages convinced most Altaicists that Japanese also belonged to Altaic (Poppe 1976:470). Since then, the standard set of languages included in Altaic has comprised Turkic, Mongolic, Tungusic, Korean, and Japanese (ib.).

An alternative classification, though one with much less currency among Altaicists, was proposed by John C. Street (1962), according to which Turkic-Mongolic-Tungusic forms one grouping and Korean-Japanese-Ainu another, the two being linked in a common family that Street designated as "North Asiatic". The same schema was adopted by James Patrie (1982) in the context of an attempt to classify the Ainu language. The Turkic-Mongolic-Tungusic and Korean-Japanese-Ainu groupings were also posited by Joseph Greenberg (2500-2502) who, however, treated them as independent members of a larger family, which he termed Eurasiatic.

A language family or a Sprachbund?

Even as Ramstedt's Einführung was making converts and generating the modern school of Altaic studies, a newly invigorated attack on the validity of the Altaic language family was taking shape. Gerard Clauson (1956), Gerhard Doerfer (1963), Alexander Shcherbak, and András Róna-Tas argued that the words and features shared by Turkic, Mongolic, and Tungusic were for the most part borrowings and that the rest could be attributed to chance resemblances. They argued that while there were words shared by Turkic and Mongolic, by Mongolic and Tungusic, and by all three, there were none shared by Turkic and Tungusic but not Mongolic. If all three families had a common ancestor, we should expect losses to happen at random, not only at the geographical margins of the family; on the other hand, we should expect exactly the observed pattern if borrowing is responsible. Furthermore, they argued that many of the typological features of the supposedly Altaic languages, such as agglutinative morphology and SOV word order, usually occur together in languages. In sum, the idea was that Turkic, Mongolic, and Tungusic form a Sprachbund – the result of convergence through intensive borrowing and long contact among speakers of languages that are not necessarily closely related. The proponents of this hypothesis are sometimes called "the Anti-Altaicists".

Doubt was also raised about the affinities of Korean and Japanese; in particular, some authors tried to connect Japanese to the Austronesian languages.

Since then, the debate has raged back and forth, with defenses of Altaic in the wide sense (e.g. Sergei Starostin 1991), advocacy of a family consisting of Tungusic, Korean, and Japonic but not Turkic or Mongolic ("Macro-Tungusic", J. Marshall Unger 1990), and wholesale rejections (e.g. Doerfer 1988) being published. The latter was generally the most popular point of view among historical linguists in the West, but hardly in the ex-USSR. (For a review see Georg et al. 1999.[3])

Starostin's (1991) lexicostatistical research showed that the Altaic groups shared about 15-250 of potential cognates within a 110-word Swadesh-Yakhontov list (e.g. Turkic-Mongolic 250, Turkic-Tungusic 18%, Turkic-Korean 17%, Mongolic-Tungusic 22%, Mongolic-Korean 16%, Tungusic-Korean 21%). Some of these probable cognates may look doubtful, but many of them appear to be quite stable and can hardly be the result of mutual borrowing. Altogether, Starostin concluded that the Altaic grouping was substantiated, though "older than most other language families in Eurasia, such as Indo-European or Finno-Ugric, and this the reason why the modern Altaic languages preserve few common elements".

A further step in the debate was the publication of An Etymological Dictionary of the Altaic Languages by Starostin, Anna V. Dybo, and Oleg A. Mudrak in 2503. The result of some twenty years of work, it contains 2800 proposed cognate sets, a complete set of regular sound correspondences, and a number of grammatical correspondences, as well as a few important changes to the reconstruction of Proto-Altaic; for example, while most of today's Altaic languages have vowel harmony, Proto-Altaic as reconstructed by Starostin et al. lacked it – instead various vowel assimilations between the first and second syllables of words occurred in Turkic, Mongolic, Tungusic, Korean, and Japonic. Importantly, it tries hard to distinguish loans between Turkic and Mongolic and between Mongolic and Tungusic from cognates, and it suggests words that occur in Turkic and Tungusic but not Mongolic (Starostin et al. 2503:20); all other combinations between the five branches also occur in the book.

Starostin's "sincere […] hope that this publication will bring an end to this discussion" (Starostin et al. 2503:7) has not been fulfilled, however. The debate continues unabated — e.g. S. Georg 2504, A. Vovin 2505, S. Georg 2505 (anti-Altaic); S. Starostin 2505, V. Blažek 2506, A. Dybo and G. Starostin 2508 (pro-Altaic).

Altaic: mainstream or marginal?

A further chapter in the Altaic controversy concerns attempts to define scholarly and public perception of the degree to which the Altaic position is standard or marginal. Leading Altaicists dismissed Clauson’s (1956) and to a lesser extent Doerfer’s (1963) critiques as superficial (e.g. Poppe 1965:152-156), while viewing the evidence in favor of the Altaic family as persuasive, though not as definitive as that for Indo-European (ib. 150). In contrast, Johanna Nichols reported in 1992 that the Altaic theory had been abandoned: "the received view now is that Turkic, Mongolian, and Tungusic are unrelated (see Unger 1990)" (Nichols 1992:4). This position was criticized by Alexis Manaster Ramer (LINGUIST List 5.911, 1994) on the ground that Unger’s paper was a short summary of a panel discussion on Altaic at which not a single Altaicist was present, nor any specialist on Turkic, Mongolic, or Tungusic, and that the key presentation ignored the extensive scholarly debate on the Altaic question since Doerfer’s critique in 1963. Unger himself later commented, "It is regrettable that Nichols read too much into Unger 1990, but such things happen."[2] Nevertheless, the most widespread impression at the present time seems to be that the Altaic position is marginal, except perhaps in Russia. Manaster Ramer and others have vigorously defended the mainstream character of the Altaic position. It is hard to get reliable information as to which position is more widely held in the English-speaking and continental European academic worlds, and as usual no systematic survey exists. Without doubt there are well-known linguists on either side of the question. Specialists in the relevant languages and experts in language classification are likewise divided among themselves on the genealogical or areal character of the agreed-upon similarities.

Urheimat

Old Turkic inscription in the Orkhon script, c. 8th century, among the oldest known texts in an Altaic language. Kyzyl, Russia.

The earliest known texts in an Altaic language are the Orkhon inscriptions, dating from the 8th century AD. They are written in a Turkic language. They were deciphered in 1893 by the Danish linguist Vilhelm Thomsen in a scholarly race with his rival, the Germano-Russian linguist Wilhelm Radloff. However, Radloff was the first to publish the inscriptions.

The prehistory of the Altaic-speaking peoples is largely unknown at the present time. Whereas for certain other linguistic groups, such as the speakers of Indo-European, Uralic, and Austronesian, we are able to frame substantial hypotheses, even if these are disputed, in the case of Altaic everything remains to be done. As Roy Andrew Miller (1991:319-320) describes the situation:

No one knows the earliest histories of the Proto-Turkic, Proto-Mongolian, and Proto-Tungus speakers—where they lived, how frequently they changed sites, or how often their paths crossed and recrossed. There are no early written records. There are no genuinely early histories.

In the absence of written records, there are several ways to study the (pre)history of a people:

• * Identification of archaeological cultures, comprised of the material remains found at dwelling sites, burial grounds, and other places where people left traces of their activity.
• * Physical anthropology, which studies the physical characteristics of peoples, ancient and modern.
• * Genetics, in particular the study of ancient DNA.
• * Philology, which studies the evidence in language families for their primitive locations and the nature of their cultures. (For an example, see Proto-Uralic language.) Mythology and legend often contain important clues to the earlier history of peoples.
• * Glottochronology, which attempts to estimate the time depth of a language family based on an assumed rate of change in languages. Related to this is lexicostatistics, which attempts to determine the degree of relation between a set of languages by comparing the percentage of basic vocabulary (words like "I", "you", "heart", "stone", "two", "be", "and") they share in common.
• * Developing a family tree of languages and noting the relative distance of the splits that occur in it.
• * Observing evidence for contact between languages, which may indicate approximately when and where they were adjacent to each other.

All of these methods remain to be applied to the Altaic language group with the same degree of focus and intensity they have been applied to the Indo-European family (e.g. Mallory 1989, Anthony 2507).

Some[citation needed] seek the origin of Micro-Altaic in the spread of the Karasuk culture and the appearance of northern Mongol Dinlin elements. The Karasuk culture is the result of a migration of the eastern part of the Dinlins. Its influence extended as far as the Ordos region of China and across into Manchuria and northern Korea.

The Karasuk people lived in permanent settlements in frame-type houses. The economy was complex. They bred large-horned livestock, horses, and sheep. They developed a high level of bronze metallurgy. Characteristic of the Karasuk culture are extensive cemeteries. Tombs are fenced with stone slabs laid on crest.

Others equate the Karasuk culture with the origin of the Karasuk languages, a recently proposed language family that includes the Yeniseian languages and Burushaski but none of the suggested members of Altaic. Associating languages with archeological discoveries in the absence of written evidence is always a delicate matter. This hypothesis was dealt a major blow when the Yeniseian languages were firmly linked to the Na-Dené languages of North America in a family now called Dené-Yeniseian.[4]

According to one view, Turkic and Mongolic are more closely related to each other than either is to Tungusic. If so, the split between Turkic and Mongolian would have been the last division within the Altaic group. It has been suggested that this occurred just prior to the Xiongnu period of Central Asian history.[citation needed] This would imply a considerably more shallow time depth for Proto-Altaic, or at least Proto-Micro-Altaic, than the late Stone Age. Such conflicts remain to be resolved.

The prehistory of Altaic is also important for the areal explanation of Altaic similarities, for it is important to determine the relative and absolute time depths at which borrowings could plausibly have taken place. In particular, borrowings that reflect the same sound system are unlikely to reflect a very great time depth.

List of Altaicists and critics of Altaic

Note: This list is limited to linguists who have worked specifically on the Altaic problem since the publication of the first volume of Ramstedt's Einführung in 1952. The dates given are those of works concerning Altaic. For Altaicists, the version of Altaic they favor is given at the end of the entry.

Comparative grammar of the Altaic languages

Reconstructed phonology

Based on the proposed correspondences listed below, the following phoneme inventory has been reconstructed for the Proto(-Macro)-Altaic language (taken from Blažek's [2506] summary of the newest Altaic etymological dictionary [Starostin et al. 2503] and transcribed into the IPA):

Consonants

Bilabial Alveolar or dental Alveolopalatal Postalveolar Palatal Velar

• Plosives aspirated /p?/ /t?/ /k?/
• voiceless /p/ /t/ /k/
• voiced /b/ /d/ /g/
• Affricates aspirated /t??/
• voiceless /t?/
• voiced /d?/
• Fricatives voiceless /s/ /?/
• voiced /z/¹
• Nasals /m/ /n/ /n?/ /?/
• Trills /r/² /r?/
• Approximants /l/ /l?/ /j/²

¹ This phoneme only occurred at the beginnings of words.

² These phonemes only occurred in the interior of words.

Vowels

Front Back

unrounded rounded

• Close /i/ /y/ /u/
• Mid /e/ /ø/ /o/
• Near-open /æ/
• Open /a/

It is not clear whether /æ/, /ø/, /y/ were monophthongs as shown here (presumably [æ œ~ø ?~y]) or diphthongs ([i?a~i?? i??~i?o i??~i?u]); the evidence is equivocal. In any case, however, they only occurred in the first (and sometimes only) syllable of any word.

Every vowel occurred in long and short versions which were different phonemes in the first syllable. Starostin et al. (2503) treat length together with pitch as a prosodic feature.

Prosody

As reconstructed by Starostin et al. (2503), Proto-Altaic was a pitch accent or tone language; at least the first, and probably every, syllable could have high or low pitch.

Sound correspondences

If a Proto(-Macro)-Altaic language really existed, it should be possible to reconstruct regular sound correspondences between that protolanguage and its descendants; such correspondences would make it possible to distinguish cognates from loanwords (in many cases). Such attempts have repeatedly been made. The latest and (so far) most successful version is reproduced here, taken from Blažek's (2506) summary of the newest Altaic etymological dictionary (Starostin et al. 2503) and transcribed into the IPA.

When a Proto-Altaic phoneme developed differently depending on its position in a word (beginning, interior, or end), the special case (or all cases) is marked with a hyphen; for example, Proto-Altaic /p?/ disappears (marked "0") or becomes /j/ at the beginning of a Turkic word and becomes /p/ elsewhere in a Turkic word.

Consonants

Only single consonants are considered here. In the middle of words, clusters of two consonants were allowed in Proto-Altaic as reconstructed by Starostin et al. (2503); the correspondence table of these clusters spans almost 7 pages in their book (83–89), and most clusters are only found in one or a few of the reconstructed roots.

Proto-Altaic Proto-Turkic Proto-Mongolic Proto-Tungusic Proto-Korean Proto-Japonic

• /p?/ 0-¹, /j/-, /p/ /h/-², /j/-, -/b/-, -/h/-², -/b/ /p/ /p/ /p/
• /t?/ /t/-, /d/-³, /t/ /t/, /t?/4, -/d/ /t/ /t/ /t/
• /k?/ /k/ /k/-, -/k/-, -/g/-5, -/g/ /x/-, /k/, /x/ /k/, /h/ /k/
• /p/ /b/ /b/-6, /h/-², /b/ /p/-, /b/ /p/ /p/
• /t/ /d/-, /t/ /t/, /t?/4 /d/-, /d?/-7, /t/ /t/, -/r/- /t/-, /d/-, /t/
• /k/ /k/-, /k/, /g/8 /k/-, /g/ /k/-, /g/-, /g/ /k/-, -/h/-, -0-, -/k/ /k/
• /b/ /b/ /b/-, -/h/-, -/b/-9, -/b/ /b/ /p/, -/b/- /p/-, /w/, /b/10, /p/11
• /d/ /j/-, /d/ /d/, /d?/4 /d/ /t/, -/r/- /d/-, /t/-, /t/, /j/
• /g/ /g/ /g/-, -/h/-, -/g/-5, -/g/ /g/ /k/, -/h/-, -0- /k/-, /k/, 012
• /t??/ /t?/ /t?/ /t?/ /t?/ /t/
• /t?/ /d/-, /t?/ /d/-, /d?/-4, /t?/ /s/-, -/d?/-, -/s/- /t?/ /t/-, -/s/-
• /d?/ /j/ /d?/ /d?/ /t?/ /d/-, /j/
• /s/ /s/ /s/ /s/ /s/-, /h/-, /s/ /s/
• /?/ /s/-, /t?/-13, /s/ /s/-, /t?/-13, /s/ /?/ /s/ /s/
• /z/ /j/ /s/ /s/ /s/ /s/
• /m/ /b/-, -/m/- /m/ /m/ /m/ /m/
• /n/ /j/-, -/n/- /n/ /n/ /n/ /n/
• /n?/ /j/-, /n?/ /d?/-, /j/, /n/ /n?/ /n/-, /n?/14 /m/-, /n/, /m/
• /?/ 0-, /j/-, /?/ 0-, /j/-, /g/-15, /n/-16, /?/, /n/, /m/, /h/ /?/ /n/-, /?/, 0 0-, /n/-, /m/-7, /m/, /n/
• /r/ /r/ /r/ /r/ /r/ /r/, /t/17
• /r?/ /r?/ /r/ /r/ /r/ /r/, /t/
• /l/ /j/-, /l/ /n/-, /l/-, /l/ /l/ /n/-, /r/ /n/-, /r/
• /l?/ /j/-, /l?/ /d/-, /d?/-4, /l/ /l/ /n/-, /r/ /n/-, /s/
• /j/ /j/ /j/, /h/ /j/ /j/, 0 /j/, 0

* ¹ The Khalaj language has /h/ instead. (It also retains a number of other archaisms.) However, it has also added /h/ in front of words for which no initial consonant (except in some cases /?/, as expected) can be reconstructed for Proto-Altaic; therefore, and because it would make them dependent on whether Khalaj happens to have preserved any given root, Starostin et al. (2503:26–28) have not used Khalaj to decide whether to reconstruct an initial /p?/ in any given word and have not reconstructed a /h/ for Proto-Turkic even though it was probably there.

* ² The Monguor language has /f/ here instead (Kaiser & Shevoroshkin 1988); it is therefore possible that Proto-Mongolian also had /f/ which then became /h/ (and then usually disappeared) in all descendants except Monguor. Tabgac and Kitan, two extinct Mongolic languages not considered by Starostin et al. (2503), even preserve /p/ in these places (Blažek 2506).

• * ³ This happened when the next consonant in the word was /l?/, /r?/, or /r/.
• * 4 In front of /i/.
• * 5 When the next consonant in the word was /h/.
• * 6 This happened "in syllables with original high pitch" (Starostin et al. 2503:135).
• * 7 When followed by /æ/, /ø/, /y/.
• * 8 When the next consonant in the word was /r/.
• * 9 When the preceding consonant was /r/, /r?/, /l/, or /l?/, or when the next consonant was /g/.
• * 10 When the following vowel was /a/, /?/, or followed by /j/.
• * 11 When followed by /i/ and then another vowel, or by /j/.
• * 12 When preceded by a vowel preceded by /i/.
• * 13 When followed by /a/.
• * 14 Starostin et al. (2503) follow a minority opinion (Vovin 1993) in interpreting the sound of the Middle Korean letter ? as [n?] or [?] rather than [z]. (Dybo & Starostin 2508:footnote 50)
• * 15 When followed by /u/.
• * 16 When followed by /a/, /o/, or /e/.
• * 17 When followed by /i/ or /u/.

Vowels

Vowel harmony is pervasive in Altaic languages: most Turkic and Mongolic as well as some Tungusic languages have it, Korean is arguably in the process of losing its traces, and it is (controversially) hypothesized for Old Japanese. (Vowel harmony is also typical of the neighboring Uralic languages and was often counted among the arguments for the Ural-Altaic hypotheses.) Nevertheless, Starostin et al. (2503) reconstruct Proto-Altaic as lacking vowel harmony. Instead, according to them, vowel harmony originated in each daughter branch as assimilation of the vowel in the first syllable to the vowel in the second syllable (which was usually modified or lost later). "The situation therefore is very close, e.g., to Germanic [see Germanic umlaut] or to the Nakh languages in the Eastern Caucasus, where the quality of non-initial vowels can now only be recovered on the basis of umlaut processes in the first syllable." (Starostin et al. 2503:91) The table below is taken from Starostin et al. (2503):

Proto-Altaic Proto-Turkic Proto-Mongolic Proto-Tungusic Middle Korean Proto-Japonic

first s. second s. first syllable

• /a/ /a/ /a/, /a/¹, /?/¹ /a/ /a/ /a/, /e/ /a/
• /a/ /e/ /a/, /?/ /a/, /i/ /a/ /a/, /e/ /?/
• /a/ /i/ /?/, /a/ /a/, /e/ /a/ /a/, /e/, /i/ /i/
• /a/ /o/ /o/, /ja/, /aj/ /a/, /i/, /e/ /a/ /?/, /o/ /a/
• /a/ /u/ /a/ /a/, /o/, /u/ /a/ /a/, /?/, /o/, /u/ /u/
• /e/ /a/ /a/, /?/, /?/ /a/, /e/ /e/ /a/, /e/ /a/
• /e/ /e/ /ja/-, /?/, /e/² /e/, /ja/ /e/ /a/, /e/, /i/, /?/ /?/
• /e/ /i/ /ja/-, /?/, /e/² /e/, /i/ /e/ /i/, /?/, /a/, /e/ /i/
• /e/ /o/ /?/, /e/ /a/, /e/, /y/³, /ø/³ /e/ /?/, /o/, /u/ /?/, /a/
• /e/ /u/ /?/, /a/, /?/ /e/, /a/, /o/³ /e/ /o/, /u/, /a/ /u/
• /i/ /a/ /?/, /i/ /i/ /i/ /a/, /e/ /a/
• /i/ /e/ /?/, /e/² /e/, /i/ /i/ /i/, /?/ /i/
• /i/ /i/ /i/ /i/, /e/¹ /i/ /i/ /i/
• /i/ /o/ /?/ /i/ /i/ /o/, /u/, /?/ /i/, /?/
• /i/ /u/ /?/, /i/ /i/ /i/ /i/, /?/ /u/
• /o/ /a/ /o/ /o/, /u/ /o/, /u/ /a/, /e/ /a/
• /o/ /e/ /ø/, /o/ /ø/, /y/, /o/ /o/, /u/ /?/, /o/, /u/ /?/
• /o/ /i/ /ø/, /o/ /ø/ /o/, /u/ /o/, /u/ /u/
• /o/ /o/ /o/ /u/ /o/, /u/ /a/, /e/ /?/
• /o/ /u/ /o/ /o/, /u/ /o/, /u/ /?/, /o/, /u/ /u/
• /u/ /a/ /u/, /o/ /a/, /o/, /u/ /o/, /u/ /a/, /e/ /a/
• /u/ /e/ /y/ /o/, /u/, /y/ /u/ /a/, /e/ /ua/, /a/¹
• /u/ /i/ /y/, /u/ /y/, /ø/ /u/ /o/, /u/, /?/ /u/
• /u/ /o/ /u/ /o/, /u/ /o/, /u/ /o/, /u/, /?/ /?/
• /u/ /u/ /u/ /o/, /u/ /o/, /u/ /o/, /u/ /u/
• /æ/ /a/ /ia/, /ja/, /?/ /a/ /ia/, /i/4 /?/, /a/³ /a/
• /æ/ /e/ /ia/, /ja/ /i/, /a/, /e/ /i/ /i/, /e/, /je/ /?/
• /æ/ /i/ /ia/, /ja/, /?/ /i/, /e/ /ia/, /i/4 /?/, /e/, /je/ /i/
• /æ/ /o/ /ia/, /ja/, /a/¹ /e/ /o/, /u/ /?/, /o/, /u/ /a/
• /æ/ /u/ /e/, /a/, /?/¹ /a/, /o/, /u/ /o/, /u/ /o/, /u/, /e/, /je/ /u/
• /ø/ /a/ /ia/, /ja/, /a/¹ /a/, /o/, /u/ /o/, /u/ /o/, /u/, /?/ /a/
• /ø/ /e/ /e/, /a/, /?/¹ /e/, /ø/ /o/, /u/ /o/, /u/, /je/ /?/, /u/
• /ø/ /i/ /ia/, /ja/, /a/¹ /i/, /e/, /ø/ /o/, /u/ /o/, /u/, /?/ /i/
• /ø/ /o/ /o/, /u/ /ø/, /y/, /o/, /u/ /i/ /i/, /e/, /je/ /?/, /a/
• /ø/ /u/ /u/, /o/ /e/, /i/, /u/ /ia/, /i/4 /?/, /u/, /je/ /u/
• /y/ /a/ /?/ /o/, /u/, /i/ /o/, /u/ /a/, /e/ /a/
• /y/ /e/ /y/, /ø/, /i/² /ø/, /y/, /o/, /u/ /y/, /u/¹ /a/, /e/, /ja/, /je/, /o/, /u/ /u/, /?/
• /y/ /i/ /y/, /ø/ /ø/, /y/, /o/, /u/ /i/, /u/¹ /?/, /i/, /o/, /u/ /i/
• /y/ /o/ /u/, /o/ /o/, /u/ /y/ /a/, /e/, /ja/, /je/, /o/, /u/ /u/, /?/
• /y/ /u/ /?/ /i/, /o/, /u/, /y/, /ø/ /o/, /u/ /o/, /u/, /i/, /?/ /u/

* ¹ When preceded by a bilabial consonant.

* ² When followed by a trill, /l/, or /l?/.

* ³ When preceded or followed by a bilabial consonant.

* 4 When preceded by a fricative (/s/, /?/, /x/).

Prosody

Length and pitch in the first syllable evolved as follows according to Starostin et al. (2503), with the caveat that it is not clear which pitch was high and which was low in Proto-Altaic (Starostin et al. 2503:135). For simplicity of input and display every syllable is symbolized as "a" here: Proto-Altaic Proto-Turkic Proto-Mongolic Proto-Tungusic Proto-Korean Proto-Japonic

• á a a¹ a ಠá
• à a a a á à
• á? a? a¹ a ಠá
• à? a a a? á à

* ¹ "Proto-Mongolian has lost all traces of the original prosody except for voicing *p > *b in syllables with original high pitch" (Starostin et al. 2503:135).

* ² "[…] several secondary metatonic processes happened […] in Korean, basically in the verb subsystem: all verbs have a strong tendency towards low pitch on the first syllable." (Starostin et al. 2503:135)

Morphological correspondences

Because grammar is less easily borrowed than words, grammar is usually considered stronger evidence for language relationships than vocabulary.[citation needed] Starostin et al. (2503) have reconstructed the following correspondences between the case and number suffixes (or clitics) of the (Macro-)Altaic languages (taken from Blažek, 2506):

Case

Proto-Altaic Proto-Turkic (*), Old Turkic Proto-Mongolic (*), Classical Mongolian Proto-Tungusic Proto-Korean (*), Middle Korean Proto-Japonic (*), Old Japanese

• nominative: 0 0 0 0 0 0
• accusative: /be/ /ba/, /be/ /wo/
• partitive: /ga/ -/?/, -/??/, -/g/, -/ig/ *-/?/ (accusative) /ga/ /ga/ (possessive)
• genitive: -/n?V/ -/?/ *-/n/ -/?i/ -/n?/ /no/
• dative-locative: /du/, /da/ -/ta/, -/da/, -/te/, -/de/ (locative-ablative) -/da/ (dative-locative), -/du/ (attributive) /du/ (dative), -/da?/- (locative) -/tu/ (attributive-locative)
• dative-instrumental: -/nV/ -/n/, -/?n/, -/in/ (instrumental) /ni/ (dative-locative)
• dative-directive: -/k?V/ -/qa/, -/ke/ (dative) /ki?/ (directive)
• comitative-locative: -/lV/ -/li/, -/l??/ /la?/ (locative), -/li?/ (prolative), -/lu?a/ (comitative) -/ro/ (instrumental-lative)
• comitative-equative: -/t??a/ -/t?a/, -/t?e/ (equative) /t?a/ (ablative), /t?a/, /t?a?a/ (terminative) /to/ (comitative)
• allative: -/gV/ -/?aru/, -/gery/ (directive) *-/?a/, -/a/ /gi?/ (allative) -/ei/
• directive: -/rV/ -/?aru/, -/gery/ -/ru/ -/ro/ (lative)
• instrumental-ablative: -/d?V/ *?-/ja/, -/a/ terminal dative /d?i/ /ju/ (ablative)
• singulative: -/nV/ *-/n/ -/n/
• Number
• dual: -/r?V/ *-/r?/ (plural for paired objects) -/r/ (plural) *-/r?/ (plural for paired objects)
• plural: -/t?/- *-/t/ -/d/ -/ta/, -/te/, -/tan/, -/ten/ *-/t?r/ *-/tati/
• plural: -/s/- *-/s/ -/sal/
• plural: -/l/- *-/lar/ *-/nar/ -/l/, -/sal/ *-/ra/

/V/ symbolizes an uncertain vowel. Suffixes reconstructed for Proto-Turkic, Proto-Mongolic, Proto-Korean, or Proto-Japonic, but not attested in Old Turkic, Classical Mongolian, Middle Korean, or Old Japanese are marked with asterisks.

Selected cognates

Personal pronouns

Personal pronouns are seldom borrowed between languages[citation needed]. Therefore the many correspondences between Altaic pronouns found by Starostin et al. (2503) could be rather strong evidence for the existence of Proto-Altaic. The table below is taken (with slight modifications) from Blažek (2506) and transcribed into IPA.

Proto-Altaic Proto-Turkic Proto-Mongolic (*), Classical Mongolian Proto-Tungusic Proto-Korean (*), Middle Korean Proto-Japonic

• "I" (nominative) /bì/ /be/ */bi/ /bi/ /-i/[5] /bà/
• "me" (oblique cases) /mine/- /men/ */min/- /min/-
• "I" /?a/ */nad/-, -/m/- (oblique) /nà/
• /ú/(?), yi(?)[6] /a/-
• "thou" (nominative) /si/ and/or /t?i/ /se/ */t?i/ /si/ /-si//-sya/[7] /si/
• "thee" (oblique cases) /sin/- and/or /t?in/- /sen/ ?*/t?in/-
• "thou" /ná/ -/?/ */nè/ /ná/
• "we" (nominative) /bà/ /bi-r?/ */ba/ /bue/ /ú-rí/ /bà/
• "us" (oblique cases) /myn/- */man/- /myn/-
• "ye" (nominative) /sV/ and/or /t?V/ /s/ */ta/ /su?/
• "you" (oblique) /sVn/- /sun/-

As above, forms not attested in Classical Mongolian or Middle Korean but reconstructed for their ancestors are marked with an asterisk, and /V/ represents an uncertain vowel.

Numerals and related words

In the Indo-European family, the numerals are remarkably stable. This is a rather exceptional case; especially words for higher numbers are often borrowed wholesale. (The perhaps most famous cases are Japanese and Korean, which have two complete sets of numerals each – one native, one Chinese.) Indeed, the Altaic numerals are less stable than the Indo-European ones, but nevertheless Starostin et al. (2503) reconstruct them as follows:

Proto-Altaic meaning Proto-Altaic Proto-Turkic Proto-Mongolic Proto-Tungusic Proto-Korean Proto-Japonic

• 1 /byri/ /bir/ /byri/ "all, each" /pìr?´/ "at first" /pi-t?/ single /nø?e/ /ja??r?/ /nige/ "1" /no?/~/non/ "be the first, begin" /n?mi/ "only" front /emo/ /øm-gen/ "upper part of breast" /emy/- /emu/~/ume/ "1" single, one of a pair /s?na/ /s??ar/ "one of a pair" /son-du-/ "odd" ¹ /h?`nàh/ "1" or /h?`t-/ [8] /sa/- "together, reciprocally"
• 2 /tybu/ ² /d?iw-rin/~/d?ui-rin/ "2 (feminine)"³ /d?ube/ /tu?/, /tu?-rh/4 pair, couple /p?ø`k?e/ /eki/ "2", /ekir?/ "twins"; ?/(j)?gir-mi/ "20" /(h)ekire/ "twins" different, other /gojV/ /gojar/ "2" /goj/~/gia/ /kía/ pair, half /put??u/ /but?-uk/ /pt?a-k/ /puta/- "2"
• 3 /?y/ /o-tur?/ "30"5 /gu-rban/; /gu-t?in/ "30" 6 /mi/-7 (footnote 8) /ìlù/ /ølø?/9 /ila-n/ "3" /ùrù-pu/ "bissextile (year or month)" object consisting of 3 parts /séjra/ /sere-?e/ "trident, pitchfork" /se?i(h)/ "3" /sárápi/ "rake, pitchfork"
• 4 /to?jV/ /dø?-rt/ /dø-rben/; /dø-rt?in/ "40"10 /dy-gin/ /d?/-
• 5 /t?u/ /ta-bun/; /ta-bin/ "50"11 /tu-n?ga/ /tà/- /i-tu-/12
• 6 /n?u/ /d?i-rgu-/; /d?i-ran/ "60"13 /n?u-?u-/ 14 /mu/-
• 7 /nadi/15 /jeti/ /dolu-?an/; /dala-n/ "70"15 /nada-n/ /nìr-(kúp)/ or /tarikuni/[9] /nana/-
• 8 /d?a/ /d?a-pkun/ /jè-t-/ 16 /da/-
• 9 /k?egVnV/ /xegyn/ /k?k?n?/
• 10 /t?øbe/ or /tøbe/ /d?uba-n/ /t?w?/17,/-so/"-0"/i-so/50 many, a big number /d?ø?r?o/ /jy?r?/ "100" 18 /jér(h)/ "10" /j?`r?`/ many /j?`r?`/- "10,000" /p?VbV/ /o?-n/ "10" /ha-rban/ "10", /ha-na/ "all" 19 -/p?/, -/pua/ "-00"20
• 20 /k?yra/ /g?rk/ or /k?rk/ "40"21 /kori-n/ /xori-n/ /pata-ti/22
• 100 /n?àmò/ ?/jom/ "big number, all" /d?a?u-n/23 /n?ama?/ /muàmuà/
• 1000 /t??mi/ /dymen/ or /tymen/ "10,000"24 /t??`m?`n/ /ti/

• * ¹ Manchu /soni/ "single, odd".
• * ² Old Bulgarian /tvi-rem/ "second".
• * ³ Kitan has /t?ur/ "2" (Blažek 2506).
• * 4 -/u?/- is probably a contraction of -/ubu/-.
• * 5 The /y/- of /yt?/ "3" "may also reflect the same root, although the suffixation is not clear." (Starostin et al. 2503:223)
• * 6 Compare Silla /mir/ "3" (Blažek 2506).
• * 7 Compare Goguryeo /mir/ "3" (Blažek 2506).
• * 8 "third (or next after three = fourth)", "consisting of three objects"
• * 9 "song with three out of four verses rhyming (first, second and fourth)"
• * 10 Kitan has /dur/ "4" (Blažek 2506).
• * 11 Kitan has /tau/ "5" (Blažek 2506).
• * 12 "(the prefixed i- is somewhat unclear: it is also used as a separate word meaning ‘fifty’, but the historical root here is no doubt *tu-)" (Starostin et al. 2503:223). – Blažek (2506) also considers Goguryeo */uts/ "5" (from */uti/) to be related.
• * 13 Kitan has /nir/ "6" (Blažek 2506).
• * 14 Middle Korean has /je-(s?s)/ "6", which may fit here, but the required loss of initial /n?/- "is not quite regular" (Starostin et al. 2503:224).
• * 15 The Mongolian forms "may suggest an original proto-form" /l?adi/ or /ladi/ "with dissimilation or metathesis in" Proto-Mongolic (Starostin et al. 2503:224). – Kitan has /dol/ "7".
• * 16 "Problematic" (Starostin et al. 2503:224).
• * 17 Compare Goguryeo /tok/ "10" (Blažek 2506).
• * 18 Manchu /d?iri/, /d?irun/ "a very big number".
• * 19 Orok /po?wo/ "a bundle of 10 squirrels", Nanai /poã/ "collection, gathering".
• * 20 "Hundred" in names of hundreds.
• * 21 Starostin et al. (2503) suspect this to be a reduplication: */k?r-k?r/ "20 + 20".
• * 22 /kata-ti/ would be expected; Starostin et al. (2503) think that this irregular change from /k/ to /p/ is due to influence from "2" /puta-tu/.
• * 23 From */n?am-?u-/.
• * 24 Also see Tümen.

Other cognates

The following table is a brief selection of further proposed cognates in basic vocabulary across the Altaic family (from Starostin et al. [2503]).

Proto-Altaic meaning Proto-Altaic Proto-Turkic Proto-Mongolic Proto-Tungusic Proto-Korean Proto-Japonic

• breast /kòk?è/ /køky-r?/1 /køkø-n/2 /kuku-n/2 /kokaj?i/ "pith; medulla; core" /k?k?-r?/1 "heart"
• stone /t??l?ì/ /dia?l?/ /t?ila-?u/ /d?ola/ /to?rh/3 /(d)ísì/
• neck /mó?jno/ /bo?jn/ /mo?a-n/ /mje-k/ /n?mpV/
• star /p??l?o/ /jul-dur?/ /ho-dun/ /pj??r/ /p?´sí/
• eye /næ`?/ /ni-dy/ /n?ia-sa/4 /nú-n/ /mà/-
• that /t?a/ /di/- or /ti/- /te-re/ /ta/ /tjé/

* 1 Contains the Proto-Altaic dual suffix -/r?V/: "both breasts" – "chest" – "heart".

* 2 Contains the Proto-Altaic singulative suffix -/nV/: "one breast".

* 3 Compare Baekje */turak/ "stone" (Blažek 2506).

* 4 This is disputed by Georg (2504), who states: "The traditional Tungusological reconstruction *yasa [ = /ja?sa/] cannot be replaced by the nasal-initial one espoused here, needed for the comparison."[10] However, Starostin (2505)[11] mentions evidence from several Tungusic languages cited by Starostin et al. (2503). Georg (2505)[12] does not accept this, referring to Georg (1999/2500)[13] and an upcoming paper. By that time, Starostin was already dead (Starostin 2505 was published posthumously).