Learn more about Proto-Indo-European language
| Albanian | Anatolian | Armenian |
Baltic | Celtic | Dacian | Germanic
Greek | Indo-Iranian | Italic | Phrygian
Slavic | Thracian | Tocharian
| Albanians | Anatolians | Armenians |
Balts | Celts | Germanic peoples
Greeks | Indo-Aryans | Indo-Iranians | Iranians
Italic peoples | Slavs | Thracians | Tocharians
|Language | Society | Religion|
|Kurgan hypothesis | Anatolia|
Armenia | India | PCT
The Proto-Indo-European language (PIE) is the hypothetical common ancestor of the Indo-European languages. Although the existence of such a language has been accepted by linguists for a long time, there has been debate about many specific details.
 Discovery and reconstruction
 When was PIE spoken?
There are several competing hypotheses about when and where PIE was spoken. The only thing known for certain is that the language must have been differentiated into unconnected daughter dialects by the mid 3rd millennium BC. Mainstream estimates of the time between PIE and the earliest attested texts (ca. 19th c., see Kültepe texts) range around 1,500 to 2,500 years, with extreme proposals diverging up to another 100% on either side:
- the 3rd millennium BC (excluding the Anatolian branch) in Armenia according to the Armenian hypothesis (proposed in the context of Glottalic theory);
- the 5th millennium BC (4th excluding the Anatolian branch) in the Pontic-Caspian steppe according to the mainstream Kurgan hypothesis;
- the 6th millennium BC in India according to Koenraad Elst's Out of India model.
- the 7th millennium BC in Anatolia (the 5th, in the Balkans, excluding the Anatolian branch) according to Colin Renfrew's Anatolian hypothesis;
- the 7th millennium BC (6th excluding the Anatolian branch) according to a 2003 glottochronological study<ref>Russell D. Gray and Quentin D. Atkinson, Language-tree divergence times support the Anatolian theory of Indo-European origin, Nature 426 (27 November 2003) 435-439</ref>
- before the 10th millennium BC, in the Paleolithic Continuity Theory
The classical phase of Indo-European comparative linguistics leads from Franz Bopp's Comparative Grammar (1833) to August Schleicher's 1861 Compendium and up to Karl Brugmann's Grundriss published from the 1880s. Brugmann's junggrammatische re-evaluation of the field and Ferdinand de Saussure's development of the laryngeal theory may be considered the beginning of "contemporary" Indo-European studies.
PIE as described in the early 1900s is still generally accepted today; subsequent work is largely refinement and systematization, as well as the incorporation of new information, notably the Anatolian and Tocharian branches unknown in the 19th century.
Notably, the laryngeal theory, in its early forms discussed since the 1880s, became mainstream after Jerzy Kuryłowicz's 1927 discovery of the survival of at least some of these hypothetical phonemes in Anatolian. Julius Pokorny's Indogermanisches Etymologisches Wörterbuch (1959) gave an overview of the lexical knowledge accumulated until the early 20th century, but neglected contemporary trends of morphology and phonology, and largely ignored Anatolian and Tocharian.
The generation of Indo-Europeanists active in the last third of the 20th century (such as Calvert Watkins, Jochem Schindler and Helmut Rix) developed a better understanding of morphology and, in the wake of Kuryłowicz's 1956 Apophonie, understanding of the ablaut. From the 1960s, knowledge of Anatolian became certain enough to establish its relationship to PIE; see also Indo-Hittite.
There is no direct evidence of PIE, because it was never written. All PIE sounds and words are reconstructed from later Indo-European languages using the comparative method and the method of internal reconstruction. The asterisk is used to mark reconstructed PIE words, such as *wódr̥ 'water', *ḱwṓn 'dog', or *tréyes 'three (masculine)'. Many of the words in the modern Indo-European languages seem to have derived from such "protowords" via regular sound changes (e.g., Grimm's law).
As the Proto-Indo-European language broke up, its sound system diverged as well, according to various sound laws in the daughter languages. Notable among these are Grimm's law and Verner's law in Proto-Germanic, loss of prevocalic *p- in Proto-Celtic, reduction to h of prevocalic *s- in Proto-Greek, Brugmann's law and Bartholomae's law in Proto-Indo-Iranian, and Grassmann's law independently in both Proto-Greek and Proto-Indo-Iranian.
 Relationship to other language families
Many higher-level relationships between PIE and other language families have been proposed. But these speculative connections are highly controversial. Perhaps the most widely accepted proposal is of an Indo-Uralic family, encompassing PIE and Uralic. The evidence usually cited in favor of this is the proximity of the proposed Urheimaten of the two families, the typological similarity between the two languages, and a number of apparent shared morphemes. Frederik Kortlandt, while advocating a connection, concedes that "the gap between Uralic and Indo-European is huge", while Lyle Campbell, an authority of Uralic, denies any relationship exists.
Other proposals, further back in time (and correspondingly less accepted), model PIE as a branch of Indo-Uralic with a Caucasian substratum; link PIE and Uralic with Altaic and certain other families in Asia, such as Korean, Japanese, Chukotko-Kamchatkan and Eskimo-Aleut (representative proposals are Nostratic and Joseph Greenberg's Eurasiatic); or link some or all of these to Afro-Asiatic, Dravidian, etc., and ultimately to a single Proto-World family (nowadays mostly associated with Merritt Ruhlen). Various proposals, with varying levels of skepticism, also exist that join some subset of the putative Eurasiatic language families and/or some of the Caucasian language families, such as Uralo-Siberian, Ural-Altaic (once widely accepted but now largely discredited), Proto-Pontic, and so on.
|Fricatives||s||h₁, h₂, h₃|
|Liquids, Glides||w||r, l||j|
The table gives the most common notation in modern publications. Variant transcriptions are given below. Raised ʰ stands for aspiration. According to the glottalic theory, the "voiced stops" of the system as described above were glottalic, perhaps ejectives, while the "voiced aspirated stops" may not have been voiced.
- Proto-Celtic, Proto-Balto-Slavic, Albanian, and Proto-Iranian merged the voiced aspirated series bʰ, dʰ, ǵʰ, gʰ, gʷʰ with the plain voiced series b, d, ǵ, g, gʷ. (However, Proto-Celtic did not merge gʷʰ and gʷ into gʷ like the others did - instead, the former became gw while the latter became b).
- Proto-Germanic underwent Grimm's law, changing voiceless stops into fricatives, devoicing unaspirated voiced stops, and de-aspirating voiced aspirates.
- Grassmann's law (Tʰ-Tʰ > T-Tʰ, e.g. dʰi-dʰeh₁- > di-dʰeh₁-) and Bartholomae's law (TʰT > TTʰ, e.g. budʰ-to- > bud-dʰo-) describe the behaviour of aspirates in particular contexts in some early daughter languages.
p, b, bʰ, grouped with the cover symbol P. b was a very rare phoneme, which is one argument in favor of the glottalic theory - it seems that languages having ejective stops tend not to have an ejective labial stop p'.
The standard reconstruction identified three coronal/dental stops: t, d, dʰ. They are symbolically grouped with the cover symbol T.
Some theorists conclude that consonant clusters of the form TK would undergo a metathesis in the proto-language, resulting in Kþ, compare Hittite dagan "earth" with Greek khthōn "earth", from ǵʰðōm, from earlier *dʰǵʰoms; Hittite hartagas "monster", Greek arktos "bear" from hrkþos from earlier h₂r̥tg̑os. Both metathetized and unmetathetized forms survive in different ablaut grades of the root dʰégʷʰ "burn" (cognate to dagaz, day) in Sanskrit, dáhati "is being burnt" < dʰégʷʰ-e- and kṣā́yat "burns" < dʰgʷʰ-éh1-.
- Palatovelars, ḱ, ǵ, ǵʰ (also transcribed k', g', g'ʰ or k̑, g̑, g̑ʰ or k̂, ĝ, ĝʰ). These were [k]- or [g]-like sounds which underwent a characteristic change in the Satem languages; they were possibly palatalized velars ([kʲ], [gʲ]) in Proto-Indo-European.
- Pure velars, k, g, gʰ.
- Labiovelars, kʷ, gʷ, gʷʰ (also transcribed ku̯, gu̯, gu̯h). Raised ʷ stands for labialization, or lip-rounding accompanying the articulation of velar sounds ([kʷ] is a sound similar to English qu in queen).
The existence of the plain velars as phonemes separate from the palatovelars and labiovelars has been disputed. In most circumstances they appear to be allophones resulting from the neutralization of the other two series in particular phonetic circumstances. It is difficult to pinpoint exactly what the circumstances of the allophony are, although it is generally accepted that neutralization occurred after s and u, and often before r. Most PIE linguists believe that all three series were distinct by late Proto-Indo-European, although a minority, including Frederik Kortlandt, believe that the plain velar series was a later development of certain satem languages; this view was originally articuled by Antoine Meillet in 1894. Those who support the view of the threefold distinction in PIE cite evidence from Albanian (Holger Pedersen, KZ 36 (1900) 277-340; Norbert Jokl, Mélanges linguistiques offerts à M. Holger Pedersen (1937) 127-161) and Armenian (Vittore Pisani, Ricerche Linguistiche 1 (1950) 165ff.) that they treated plain velars differently from the labiovelars in at least some circumstances, as well as the fact that Luwian apparently has distinct reflexes of all three series: *ḱ > z (probably [ts]); *k > k; *kʷ > ku (probably [kʷ]) (Craig Melchert, Studies in Memory of Warren Cowgill (1987) 182–204). Kortlandt, however, disputes the significance of this evidence (Recent developments in historical phonology (1978) 237-243 = ). Ultimately, this dispute may be irresoluble -- analogical developments tend to quickly obscure the original distribution of allophonic variants that have been phonemicized, and the time frame is too great and the evidence too meager to make definite conclusions as to when exactly this phonemicization happened.
s (with the voiced allophone z). The "laryngeals" may have been fricatives, but there is no consensus as to their phonetic realization. There were also fricative allophones of t, d, usually transcribed þ, ð.
The symbols h₁, h₂, h₃, with cover symbol H (or ə₁, ə₂, ə₃ and ə), stand for three hypothetical "laryngeal" phonemes. There is no consensus as to what these phonemes were, but it is widely accepted that h₂ was probably uvular or pharyngeal, and that h₃ was labialized. Commonly cited possibilities are ʔ ʕ ʕʷ and x χ~ħ xʷ; there is some evidence that h₁ may have been two consonants, ʔ and h, that fell together.
 Nasals and Liquids
r, l, m, n, with vocalic allophones r̥, l̥, m̥, n̥, grouped with the cover symbol R.
w, y (also transcribed u̯, i̯) with vocalic allophones u, i.
- Short vowels a, e, i, o, u
- Long vowels ā, ē, ō; sometimes a colon (:) is employed to indicate vowel length instead of the macron sign (a:, e:, o:).
- Diphthongs ai, au, āi, āu, ei, eu, ēi, ēu, oi, ou, ōi, ōu
- vocalic allophones of consonantal phonemes: u, i, r̥, l̥, m̥, n̥.
Other long vowels may have appeared already in the proto-language by compensatory lengthening: ī, ū, r̥̄, l̥̄, m̥̄, n̥̄.
It is often suggested that all a sounds (short and long) were earlier derived from an e preceded or followed by h₂, but Mayrhofer (1986: 170 ff.) has argued that PIE did in fact have a and ā phonemes independent of h₂.
All Indo-European languages are inflected languages (although many modern Indo-European languages, including Modern English, have lost much of their inflection). By comparative reconstruction, it is quite likely that at least the latest stage of the common PIE mother languages (Late PIE) was an inflectional language, which was more suffixing than prefixing. However, by means of internal reconstruction and morphological (re-)analysis of the reconstructed, seemingly most ancient PIE word forms, it has recently been shown to be very probable that at a more distant stage PIE (Early PIE) may have been a root-inflected language, as was Proto-Semitic. As a consequence, it seems to be highly probable that PIE once was of the root-and-pattern morphological type.<ref>Pooth, Roland A. (2004): "Ablaut und autosegmentale Morphologie: Theorie der uridg. Wurzelflexion", in: Indogermanistik, Germanistik, Linguistik, Akten der Arbeitstagung der Indogermanischen Gesellschaft, Jena, 18.-20.09.2002. Ed. by M. Kozianka, R. Lühr, S. Zeilfelder. Hamburg, pp. 401-471 (ISBN 3-8300-1464-3).</ref>
Indo-European had a characteristic general ablaut sequence that contrasted the vowel phonemes o/e/Ø through the same root.
There are two major types of declension, thematic and athematic. Thematic nominal stems are formed with a suffix -o- (in vocative -e) and the stem does not undergo ablaut. The athematic stems are more archaic, and they are classified further by their ablaut behaviour (acro-dynamic, protero-dynamic, hystero-dynamic and holo-dynamic, after the positioning of the early PIE accent (dynamis) in the paradigm).
|(Beekes 1995)||(Ramat 1998)|
|Masculine and Feminine||Neuter||Masculine and Feminine||Neuter||Masculine||Neuter|
|Nominative||-s, 0||-es||-h₁(e)||-m, 0||-h₂, 0||-ih₁||-s||-es||-h₁e?||0||(coll.) -(e)h₂||-os||-ōs||-oh₁(u)?||-om|
|Accusative||-m||-ns||-ih₁||-m, 0||-h₂, 0||-ih₁||-m̥||-m̥s||-h₁e?||0||-om||-ons||-oh₁(u)?||-om|
|Genitive||-(o)s||-om||-h₁e||-(o)s||-om||-h₁e||-es, -os, -s||-ōm||-os(y)o||-ōm|
|Locative||-i, 0||-su||-h₁ou||-i, 0||-su||-h₁ou||-i, 0||-su, -si||-oi||-oisu, -oisi|
|Vocative||0||-es||-h₁(e)||-m, 0||-h₂, 0||-ih₁||-es||(coll.) -(e)h₂|
PIE pronouns are difficult to reconstruct due to their variety in later languages. This is especially the case for demonstrative pronouns.
PIE had personal pronouns in the first and second person, but not the third person, where demonstratives were used instead. The personal pronouns had their own unique forms and endings, and some had two distinct stems; this is most obvious in the first person singular, where the two stems are still preserved in English I and me. According to Beekes (1995), there were also two varieties for the accusative, genitive and dative cases, a stressed and an enclitic form.
|Personal pronouns (Beekes 1995)|
|First person||Second person|
|Accusative||h₁mé, h₁me||nsmé, nōs||twé||usmé, wōs|
|Genitive||h₁méne, h₁moi||ns(er)o-, nos||tewe, toi||yus(er)o-, wos|
|Dative||h₁méǵʰio, h₁moi||nsmei, ns||tébʰio, toi||usmei|
As for demonstratives, Beekes (1995) tentatively reconstructs a system with only two pronouns: so/seh₂/tod "this, that" and h₁e/ (h₁)ih₂/(h₁)id "the (just named)" (anaphoric). He also postulates three adverbial particles ḱi "here", h₂en "there" and h₂eu "away, again", from which demonstratives were constructed in various later languages.
There was also an interrogative/indefinite pronoun with the stem kʷe-/kʷi- (adjectival kʷo-), and probably a relative pronoun with the stem yo-. A third-person reflexive pronoun se (acc.), sewe, sei (gen.), sébʰio, soi (dat.), parallel to the first and second person singular personal pronouns, also existed, as well as possessive pronominal adjectives.
PIE had a separate set of endings for pronouns; many of these were later borrowed as nominal endings.
The Indo-European verb system is complex and exhibits a system of ablaut, as is still visible in the Germanic languages (among others)—for example, the vowel in the English verb to sing varies according to the conjugation of the verb: sing, sang, and sung.
Verbs have at least four moods (indicative, imperative, subjunctive and optative, as well as possibly the injunctive, reconstructible from Vedic Sanskrit), two voices (active and mediopassive), as well as three persons (first, second and third) and three numbers (singular, dual and plural). Verbs are conjugated in at least three "tenses" (present, aorist, and perfect), which actually have primarily aspectual value. Indicative forms of the imperfect and (less likely) the pluperfect may have existed. Verbs were also marked by a highly developed system of participles, one for each combination of tense and mood, and an assorted array of verbal nouns and adjectival formations.
A number of secondary forms could be created, such as the causative, intensive and desiderative; technically these were part of the derivational system rather than the inflectional system, as they existed only for certain verbs and did not necessarily have completely predictable meanings (compare the remnants of causative constructions in English – to fall vs. to fell, to sit vs. to set, to rise vs. to raise and to rear). The above-mentioned verbal nouns and adjectives were likewise part of the derivational system (compare the formation of verbal nouns in English, using -tion, -ence, -al, etc.), and it appears that the same originally applied to the different verb tenses. Some verbs in Ancient Greek still have perfect tenses with unpredictable meanings – from histēmi "I set, I cause to stand": hestēka "I am standing"; from mimnēiskō "I remind": memnēmai "I remember"; from peithō "I persuade": pepoitha "I trust" as well as pepeika "I have persuaded"; from phūō "I produce": pephūka "I am (by nature)". The present tense in Ancient Greek and in Sanskrit is formed by the unpredictable addition of one of a number of suffixes (at least 10, in Sanskrit; at least 6, in Greek) to the verbal root; the aorist and perfect are likewise formed, in each case from their own set of suffixes (7 for the Sanskrit aorist, at least 3 for the Greek aorist), with little or no relation between the suffixes used in one tense and in another. (The perfect tense in Latin is likewise unpredictable, formed in one of at least six ways.) Sometimes more than one suffix can be applied to the same root, producing different present, aorist and/or perfect stems for the same verb, sometimes with the same meaning, sometimes with different meanings (see the above example with the Greek verb peithō). All of this suggests that the various tenses were originally independent lexical formations, similarly to the way that verbal nouns in English are formed unpredictably in English from different suffixes, sometimes with two or more formations that may differ in meaning: reference vs. referral, transference vs. transferral vs. transfer, recitation vs. recital, delivery vs. deliverance etc. (This is more understandable if one considers that the original meaning of these tenses was aspectual.) Only later, and gradually, were these various forms combined into a single set of inflectional paradigms. Vedic Sanskrit had still not completed the process, and even Ancient Greek has places where the old unorganized system still shows through. (As a result, verbs in Vedic Sanskrit have the appearance at first glance of a fantastically complex and disorganized system, with numerous redundancies combined with inexplicable holes. The system of PIE must have looked even more strongly like this.)
The primary distinction in verbs between the different ways of forming the present tenses was between thematic (ō) classes, with a "thematic" vowel o or e before the endings, and athematic (mi) classes, with endings added directly to the root. The endings themselves differed somewhat, at the very least in the first-person singular, with the endings as indicated (ō vs. mi). Traditional accounts say that this is the only form where the endings differed, except for the presence or absence of the thematic vowel; but some newer researchers, e.g. Beekes (1995), have proposed a totally different set of thematic endings, based primarily on Greek and Lithuanian. These proposals are still controversial, however.
|Buck 1933||Beekes 1995|
The original meanings of the past tenses (aorist, perfect and imperfect) are often assumed to match their meanings in Greek. That is, the aorist represents a single action in the past, viewed as a discrete event; the imperfect represents a repeated past action or a past action viewed as extending over time, with the focus on some point in the middle of the action; and the perfect represents a present state resulting from a past action. This corresponds, approximately, to the English distinction between "I ate", "I was eating" and "I have eaten", respectively. (Note that the English "I have eaten" often has the meaning, or at least the strong implication, of "I am in the state resulting from having eaten", in other words "I am now full". Similarly, "I have sent the letter" means approximately "The letter is now (in the state of having been) sent". However, the Greek, and presumably PIE, perfect, more strongly emphasizes the state resulting from an action, rather than the action itself, and can shade into a present tense.)
Note that in Greek the difference between the present, aorist and perfect tenses when used outside of the indicative (that is, in the subjunctive, optative, imperative, infinitive and participles) is almost entirely one of grammatical aspect, not of tense. That is, the aorist refers to a simple action, the present to an ongoing action, and the perfect to a state resulting from a previous action. An aorist infinitive or imperative, for example, does not refer to a past action, and in fact for many verbs (e.g. "kill") would likely be more common than a present infinitive or imperative. (In some participial constructions, however, an aorist participle can have either a tensal or aspectual meaning.) It is assumed that this distinction of aspect was the original significance of the PIE "tenses", rather than any actual tense distinction, and that tense distinctions were originally indicated by means of adverbs, as in Chinese. However, it appears that by late PIE, the different tenses had already acquired a tensal meaning in particular contexts, as in Greek, and in later Indo-European languages this became dominant.
The meanings of the three tenses in the oldest Vedic Sanskrit, however, differs somewhat from their meanings in Greek, and thus it is not clear whether the PIE meanings corresponded exactly to the Greek meanings. In particular, the Vedic imperfect had a meaning that was close to the Greek aorist, and the Vedic aorist had a meaning that was close to the Greek perfect. Meanwhile, the Vedic perfect was often indistinguishable from a present tense (Whitney 1924). In the moods other than the indicative, the present, aorist and perfect were almost indistinguishable from each other. (The lack of semantic distinction between different grammatical forms in a literary language often indicates that some of these forms no longer existed in the spoken language of the time. In fact, in Classical Sanskrit, the subjunctive dropped out, as did all tenses of the optative and imperative other than the present; meanwhile, in the indicative the imperfect, aorist and perfect became largely interchangeable, and in later Classical Sanskrit, all three could be freely replaced by a participial construction. All of these developments appear to reflect changes in spoken Middle Indo-Aryan; among the past tenses, for example, only the aorist survived into early Middle Indo-Aryan, which was later displaced by a participial past tense.)
The numbers are generally reconstructed as follows:
|Sihler 1995, 402–24||Beekes 1995, 212–16|
|three||*trei- (full grade)/*tri- (zero grade)||*treies|
|four||*kʷetwor- (o-grade)/*kʷetur- (zero grade), |
see also the kʷetwóres rule
|six||*s(w)eḱs; originally perhaps *weḱs||*(s)uéks|
|eight||*oḱtō, *oḱtou or *h₃eḱtō, *h₃eḱtou||*h₃eḱteh₃|
|twenty||*wīḱm̥t-; originally perhaps *widḱomt-||*duidḱmti|
|thirty||*trīḱomt-; originally perhaps *tridḱomt-||*trih₂dḱomth₂|
|forty||*kʷetwr̥̄ḱomt-; originally perhaps *kʷetwr̥dḱomt-||*kʷeturdḱomth₂|
|fifty||*penkʷēḱomt-; originally perhaps *penkʷedḱomt-||*penkʷedḱomth₂|
|sixty||*s(w)eḱsḱomt-; originally perhaps *weḱsdḱomt-||*ueksdḱomth₂|
|seventy||*septm̥̄ḱomt-; originally perhaps *septm̥dḱomt-||*septmdḱomth₂|
|eighty||*oḱtō(u)ḱomt-; originally perhaps *h₃eḱto(u)dḱomt-||*h₃eḱth₃dḱomth₂|
|ninety||*(h₁)newn̥̄ḱomt-; originally perhaps *h₁newn̥dḱomt-||*h₁neundḱomth₂|
|hundred||*ḱm̥tom; originally perhaps *dḱm̥tom||*dḱmtóm|
Lehmann (1993, 252-255) believes that the numbers greater than ten were constructed separately in the dialects groups and that *ḱm̥tóm originally meant "a large number" rather than specifically "one hundred."
 Sample texts
As PIE was spoken by a prehistoric society, no genuine sample texts are available, but since the 19th century modern scholars have made various attempts to compose example texts for purposes of illustration. These texts are educated guesses at best; Calvert Watkins in 1969 rightly observes that in spite of its 150 years' history, comparative linguistics is not in the position to reconstruct a single well-formed sentence in PIE. Nevertheless, such texts do have the merit of giving an impression of what a coherent utterance in PIE might have sounded like.
Published PIE sample texts:
- Schleicher's fable (Avis akvasas ka) by August Schleicher (1868), modernized by Hermann Hirt (1939) and Winfred Lehmann and Ladislav Zgusta (1979)
- The king and the god (rēḱs deiwos-kʷe) by S. K. Sen, E. P. Hamp et al. (1994)
- Beekes, Robert S. P. (1995). Comparative Indo-European Linguistics: An Introduction. Amsterdam: John Benjamins. ISBN 90-272-2150-2 (Europe), ISBN 1-55619-504-4 (U.S.).
- Buck, Carl Darling (1933). Comparative Grammar of Greek and Latin. Chicago: University of Chicago Press. ISBN 0-226-07931-7.
- Lehmann, W., and L. Zgusta. 1979. Schleicher's tale after a century. In Festschrift for Oswald Szemerényi on the Occasion of his 65th Birthday, ed. B. Brogyanyi, 455–66. Amsterdam.
- Mayrhofer, Manfred (1986). Indogermanische Grammatik, i/2: Lautlehre. Heidelberg: Winter.
- Sihler, Andrew L. (1995). New Comparative Grammar of Greek and Latin. Oxford: Oxford University Press. ISBN 0-19-508345-8.
- Szemerényi, Oswald (1996). Introduction to Indo-European Linguistics. Oxford.
- Whitney, William Dwight (1924). Comparative Grammar of Greek and Latin. Delhi: Motilal Banarsidass Publishers Private Limited (reprint). ISBN 81-208-0621-2 (India), ISBN 0-486-43136-3 (Dover, US).
 See also
- Indo-European studies
- Proto-Indo-European religion
- List of Indo-European roots
- Proto-World language
- Indo-European s-mobile
 External links
- American Heritage Dictionary:
- PIE grammar
- Indo-European Etymological Dictionary (Leiden University)
- Indo-European Documentation Center at the University of Texas
- "The Indo-Uralic Verb" by Frederik Kortlandt
- Say something in Proto-Indo-European (by Geoffrey Sampson)
- An Overview of the Proto-Indo-European Verb System (by Piotr Gąsiorowski)
- Many PIE example texts
- PIE root etymology database, compiled from Walde-Pokorny by L. Nikolayev and from Friedrich, Tischler (Hittite) and Adams (Tocharian) by Sergei Starostin.
- On the internal classification of Indo-European languages: survey by Václav Blažek. Linguistica ONLINE. ISSN 1801-5336 (Brno, Czech Republic)bg:Индоевропейски праезик
br:Indez-europeg ca:Protoindoeuropeu de:Indogermanische Ursprache es:Protoindoeuropeo fi:Indoeurooppalainen kantakieli hu:Indoeurópai alapnyelv fr:Indo-européen commun it:Indoeuropeo ja:インド・ヨーロッパ祖語 la:Lingua prisca Indoeuropaea nl:Proto-Indo-Europees no:Proto-indoeuropeisk språk pl:Język praindoeuropejski pt:Proto-Indo-Europeu ro:Limba proto-indo-europeană ru:Праиндоевропейский язык sv:Urindoeuropeiska zh:原始印歐語