Indo-European original language

from Wikipedia, the free encyclopedia

The Indo-European original language (or: Indo-European basic language or Ur-Indo-European ) is the common forerunner ( original language ) of the Indo-European languages that has not been documented but has been developed through linguistic methods . For this language family, the term “Indo-European” is also common (especially internationally), and accordingly the original language is then also called Proto-Indo-European (PIE). The term “Indo-European” is meant to mean that the language family occurs in an area between the Germanic distribution area and India; in fact, however, most of the “Indo-European” languages ​​are neither Germanic nor Indian, and the original language has no special connection with the Germanic or Indian subsidiary languages ​​in particular. Likewise, the term Indo-European should not mean that this original language must necessarily have originated in Europe (see also in the article Indo-European languages ​​# The term ).

It is one of the important achievements of linguists since the beginning of the 19th century to have plausibly reconstructed the vocabulary and grammatical structure of this original language from the investigation of the similarities and the systematic differences of the Indo-European languages . The reconstruction is based primarily on similarities in grammatical forms and on related words ( cognates ). A high number of cognates indicates a relationship if the vocabulary to be compared comes from the basic vocabulary .

Dating and localization of the Urindo-European

Image 1: Possible spread of Urindo-European around 3500 BC With splitting off of the Anatolian languages . (. See Kurgan hypothesis .)

Jamna culture : a Kurgankultur , possible speakers of Indo-European proto-language;
Cucuteni-Tripolje culture : a culture that succeeds the ceramic band culture ;
Vinča culture : named after the place where Vinča was found near Belgrade ;
Maykop culture : named after the city of Maykop in Russia.
Distribution area and geographic extent of the Yamnaja culture (3200–2300 BC).

Surrounding cultures: Dimple ceramic culture (until 2000 BC); Kama culture (5000-3000 BC); Celteminar culture (5500-3500 BC); Maykop culture (4000 and 3200 BC).

Due to the common vocabulary of the subsequent languages, which includes, for example, the words for “ wheel ”, “ axle ” and other important terms in wagon technology (cf. also the chapter on vocabulary analysis ), most researchers do not assume that the language was split before 3400 BC. Chr. From. Archeology dates the first secure use of bicycles to this time , also in the assumed language area. The degree of diversity in language monuments from the second millennium BC onwards. Evidence of subsequent conversations suggests a separation time after around 3000 BC. No longer appear plausible.

The geographical and temporal classifications of this language are uncertain. The representation shown in the map (Fig. 1) is considered possible by experts - but other areas have also been suggested.

Similarities in the follow-up conversations

Since the original Indo-European language has not been passed down directly, all sounds and words were made accessible using the comparative method ( language reconstruction ). Many words in today's Indo-European languages ​​are derived from these original words through regular sound changes . This is much clearer in earlier forms of these languages. The grammatical structures of the languages ​​also show great similarities (especially in the older language levels). After researchers like Franz Bopp and Jacob Grimm had presented the similarities in detail in the first half of the 19th century , August Schleicher tried in 1861 to reconstruct the assumed common root. Since then and to this day, this reconstruction has been continuously revised on the basis of new discoveries and analyzes.

Following August Schleicher, reconstructed forms are marked with an asterisk ( asterisk ): * wódr̥ 'water', * ḱwṓ (n) 'dog' or * tréyes 'three'. The following table should serve as a first illustration of the similarities, showing some numerals in various subsequent languages ​​and in the Indo-European reconstruction. (The spelling of the reconstructed words is explained below .)

number Hittite Greek Vedic Avestisch Latin Welsh Gothic Armenian Tocharian A Old Church
Lithuanian Kurdish Indo-European
1   heîs (< * hens < * sems ) éka aēuua ūnus ( older oinos ) U.N ains mi sas inŭ vienas yek * oyno-, * oyko-, * sem-
2 Dan dýō (epic) dvā́ duua duo last twai Erkow wu dŭva you you * d (u) wóh₁
3 teri- trip tráyas θrāiiō trēs tri þreis erekʿ tre trije trỹs * tréyes
4th meya- téttares (Attic) catvā́ras caθuuārō quattuor pedwar fidwor čʿorkʿ śtwar četyre keturì çwar * kʷetwóres
5   pénte páñca panca quinque pump fimf hung päñ pętĭ penkì pênc * pénkʷe
6th   héx ṣáṣ xšuuaš sex change saíhs vecʿ ṣäk šestĭ šešì şeş * swék̑s
7th sipta- heptá saptá hapta septem saith sibun ewtʿn late sedmĭ septynì hewt * septḿ̥
8th   oct aṣṭā́ ašta octō wyth ahtau owtʿ ok osmĭ aštuonì heşt * ok̑tṓ
9   ennéa náva nauua novem naw niun inn ñu devętĭ devynì * néwn̥
10   déka the A the A decem deg taíhun tasn śäk desętĭ dẽšimt de * dék̑m̥
20th   wíkati ( Doric ) vimśatí vīsaiti vīgintī ugain ( ugain older ) - kʿsan wiki - - bîst * wi / ī- (d) k̑m̥-t-ī́
100   hekatón śatám satəm centum cant dog - känt sŭto šim̃tas sed * (d) k̑m̥-tóm
Probable distribution of the language groups around 1500 BC Chr.
(Urnenf. = Urnfield culture ). The oasis culture is believed by many to be the carrier of the original Indo-Iranian.
Hypothetical distribution of Indo-European languages ​​2500 BC Chr.

Not only word equations but also grammatical structures show such great similarities in the Indo-European languages ​​that one has to start from a common origin of these languages. The counter-model of a linguistic union , i.e. a group of originally independent languages ​​that would have approached each other through mutual influencing, is ruled out in view of the nature of the observed phenomena.

Nevertheless, it would be a mistake to think of Urindo-European as a single language spoken in exactly the same way by a group of people. On the one hand, language elements are to be assumed that did not leave any traces in any of the subsequent conversations and therefore cannot be reconstructed; on the other hand, it must be noted that the reconstruction encompasses a spatially extended dialect continuum and a period of many centuries.

The language groups and their oldest traditions

Individual dialect families separated from the continuum of the Indo-European original language at different times. The linguistic isolation can be read from the characteristics of the lexicon and the morphology as well as from specific sound laws. The reconstruction of Urindo-European is based on linguistic monuments of the various Indo-European language groups. Naturally, particularly early language monuments are of primary interest.

The table gives an overview of the language groups from the perspective of dealing with the original language. More about the language groups themselves and their further development can be found in the individual articles as well as in the main article Indo-European languages .

Language group Oldest traditions Later important traditions Early distribution area Important aspects for the reconstruction
Anatolian languages Old Ethite cuneiform tablets from the 16th century BC Chr. Asian part of today's Turkey
  • Laryngeal sounds partially preserved directly .
  • Noticeable grammatical deviations from other languages, which some see as preserved archaic structures, others as innovations.
Greek language Linear-B clay tablets from the 2nd millennium BC BC , who document the Mycenaean Greek in short list-like administrative texts. Greece, western Asia Minor, southern Italy, Mediterranean area
  • Tense mode aspect system of the verb
  • Three different vocal reflexes of the laryngals
  • Ablaut.
Indo-Aryan languages The Rigveda is believed to be in India in the later 2nd millennium BC. BC originated. Purely oral transmission of the Vedic texts up to the second millennium after the Christian era, but good preservation of the language level due to the high religious priority of the unadulterated preservation of the wording.
  • Sanskrit is derived from an ancient Indian dialect related to Vedic.
  • From Panini in the 5th or 4th century BC. Chr. Grammatically fixed, but not yet written down.
  • Oldest written records: Middle Indian inscriptions by Ashoka (3rd century BC)
  • Sanskrit has been the language of education, literature and sacred since then in the form established by Panini.
North India
  • Before Schleicher's first reconstruction, Sanskrit was used in research as an approximation of the original language.
  • Voiced aspirated plosives,
  • Noun inflection,
  • Accent and ablaut classes,
  • Roots of words.
Iranian languages The Avestan , the language of religious texts of Zarathustra , is with them in the v 10th century. Chr. Dated. These texts were passed down orally and only recorded in writing in the middle of the first millennium AD.
  • Old Persian was used under Darius I in the middle of the 5th century BC. Chr. And his successors in a specially developed (but not very suitable for reproducing the language) writing system, the ancient Persian cuneiform , recorded in a few inscriptions.
Territory of today's Iran , Afghanistan , Tajikistan and Kurdistan
  • Smaller text corpus, therefore less important for the reconstruction than in the related Vedic.
  • Avestic findings as confirmation and corrective of the Vedic.
Italian languages Italian languages : Oldest Italian language monuments from the 6th century BC Chr. In Oscan , Umbrian , Faliscan etc. Large part of the area of ​​today's Italy .
  • The large corpus provides a lot of material for the word roots and the morphology.
  • Extensive innovation in syntax only allows indirect conclusions.
Celtic languages Short texts are from the time since the 2nd century BC. Chr. Handed down. Irish and Welsh language literature from the Middle Ages, e.g. Ulster cycle , Mabinogion All of Europe, from the Iberian region to Asia Minor, from the British Isles to Northern Italy, see list of Celtic tribes .
  • The discovery and the proof that Celtic is one of the Indo-European languages ​​at all is an early triumph of Indo-European studies .
Germanic languages According to names and short runic texts from the 1st century BC. . AD is Wulfila Bible translation in the 4th century to the Gothic , the oldest Germanic larger document. A number of very old Germanic words have survived in Finnish loanwords. Old High German , Old English , Old Norse , Old Saxon texts from the second half of the first millennium AD. Through the great migration across Europe and North Africa. Gothic language remnants were recorded in the Crimea in the 16th century .
  • The Germanic languages ​​were traditionally a research area that was heavily studied by the Indo-Europeanists.
  • Verner's law allows direct conclusions to be drawn about the Indo-European word accent.
Armenian language The oldest traditions begin with the creation of the Armenian script in 406   Armenia , Armenian Highlands , Eastern and South-Eastern Asia Minor
  • Similarities with Greek, Indo-Iranian and Phrygian, especially the augment .
Tocharian languages In the two Tocharian languages, mainly Buddhist texts in a form of Brahmi script from the 6th century to the 8th century have survived.   In what is now the Uighur Autonomous Region of Xinjiang in the far northwest of China
Slavic and Baltic languages The oldest surviving Slavic language is Old Church Slavonic from the 2nd half of the 9th century. The Baltic languages ​​are only passed down from the 14th century. The Baltic -Slavic hypothesis , according to which the Slavic and Baltic languages ​​go back to a common intermediate form, is widely accepted, but is contested by some researchers.    
  • Particularly conservative morphology
Albanian language The oldest surviving Albanian texts date from the 15th century. Connections with the Illyrian language cannot really be proven due to a lack of sufficient evidence.   Today's Albania and the surrounding area.
  • The Albanian vocabulary consists mainly of borrowings from ancient Greek and Latin, from Romance and South Slavic languages, as well as from a Mediterranean substrate .
Oscar inscription, 5th century BC Chr.
Bruties esum - I belong to Brutus (?); read from right to left

There are also some old individual languages ​​that have been handed down only to a limited extent, which (mostly due to a lack of material) cannot be assigned to any of the known groups, for example those in the middle of the first millennium BC. Phrygian language spoken in Greek script , then also Thracian , Macedonian , Illyrian , Venetian or Lusitanian .


Indo-European was an inflectional language . There are many indications that inflection only developed in the language over time. In the following languages, the inflection was reduced to varying degrees - only a little in the Baltic languages, most notably in English , New Persian and Afrikaans , which apart from traces of inflection have moved very close to the isolating languages .

According to W. Lehmann (1974), the word order type was SOV (i.e. the predicate was in declarative sentences at the end of the sentence) with the properties that are typically associated with it (post positions, preceding attributes and relative clauses, etc.). Other types developed in the ensuing languages: VSO in Island Celtic , SVO in Romansh.

In terms of the so-called relational typology , Indo-European (like most of the languages ​​spoken today) was an accusative language . Lehmann assumes that an earlier language level had the character of an active language . Many of the modern Indo-Aryan languages ​​(for example Hindi ) have adopted the split ergativity type .

Phonology and Phonetics

The phonemes shown below are reconstructed for the Indo-European original language . Going back to Karl Brugmann , variants of a system of Latin letters with a few superscripts and subscripts as well as diacritical marks are used for written representation.


labial alveolar palatal velar labio velar laryngeal
voiceless plosives p t k  
voiced plosives b d ǵ G G  
voiced - aspirated plosives ǵʰ G G  
Nasals m n        
Fricatives   s       h₁, h₂, h₃
Approximants w r, l y      

The nasals m and n as well as the approximants (proximity sounds) - the liquid (flowing sounds) l and r and the anguste (angel sounds) y and w - are called resonants . They have the ability to become syllabic around other consonants. Between plosives, the fricatives h₁, h₂, h₃ also become syllabic (they then appear in Greek as e, a, o , in Indoiran. As i , in Slav. As o and otherwise as a ) or disappear completely ( ved.pitā́ ‚ Father ', opposite avest. Ptā but Dat. Fədrōi ).

y was (presumably) as [⁠ j ⁠] as in German y a , w as [⁠ w ⁠] in English w ater pronounced in diphthongs ey , aw ([ EJ ], [ aʊ̯ ] English. p a per , dt. p au se ). For the phonetic realization of the palatals , ǵ and ǵʰ cf. [ ] (British English c ube ), to which the Labiovelare , and gʷʰ Italy. cinque 'five' [ ] (pronounced k with rounded lips). The voiced aspirated plosives of Indo-European do not occur in modern European languages; they are still preserved in Indian languages ​​(e.g. Hindi ).

The term "Laryngal" for the sounds designated with h₁, h₂, h₃ was chosen historically without a basis in the reconstruction. There are three unknown sounds (some researchers also suggest four or more laryngals). There are various assumptions about possible pronunciations of these sounds (see e.g. Lehmann or Meier-Brügger). The laryngeal theory was introduced into Indo-European studies by Ferdinand de Saussure in 1878 , but it took about 100 years before it was generally accepted.

The s was voiceless ( [⁠ s ⁠] ) but before voiced sounds had a voiced allophone , z. B. * ni-sd-ó- 'nest' (actually 'sitting down, sitting down'), phonetically then / nizdos / (cf. Latin nīdus , German nest ).

The so-called glottal hypothesis revises this classical reconstruction system with regard to plosives to a large extent. The revision relates essentially to the phonetics , i.e. the assumed pronunciation of the sounds; it does not change the phonological system (the relationships between the sounds) as a whole. There is no longer any voicing or aspiration; instead of voiceless - voiced - voiced aspirated occurs fortis - glottal - lenis ; the series * p * b * bʰ is then z. E.g. with * p * p '* b ( Vennemann ; Gamqrelidse and Iwanow) or with * p : * p' * p (Kortlandt). The reason for the Glottalhypothese delivered the rare occurrence of the phoneme / b /, as well as the unusual, almost unique among the languages of the world constellation aspirated voiced plosives in the absence aspirated voting loose plosive. This theory is still debated today, but it is not the majority opinion of the experts.

The reconstructs are mostly presented phonologically. The sometimes unspeakable appearing clusters of consonants suggest that the phonetics of language included scion vowels (e.g. the " Schwa secundum"), assimilations and similar phenomena.

Occurrence of the consonants

labial alveolar palatal velar labiovelar laryngeal
voiceless plosives * p ṓds (Gen. p ḗds) 'foot' * t ers- 'dry'
(see dry )
* ḗr (Gen. rd-ós) 'heart' * leu k - 'shine' * í-s, * ó-d 'who? what?'  
voiced plosives * b el 'strength' (cf. Latin dē bil is 'spiritually powerless') * d éḱm̥t 'ten' * ǵ enu-s 'knee' * h₂eu g - 'multiply'
(cf. growing )
* ne , * no 'naked'  
voiced aspirated plosives * 'bear'
(see stretcher )
* me -io- 'medium' * h₂en ǵʰ (a narrow en) * lh₂dʰ- 'smooth' * gʷʰ er-mo- 'warm'  
Nasals * m en- 'think'
(see warning )
* n eh₂-s (Gen. n h₂-s-ós) 'nose'        
Fricatives   * s ed- 'sit'       * h₂ ue h₁ - 'blow', * de h₃ - 'give' (cf. Latin dare ), * h₃ ep- 'work' (cf. Latin opus 'work', German practice )
Approximants * né w o- 'new' * p r ó 'forward', * l egʰ- 'lie' * h₂é y -es 'metal'
(see ore )

Development of the consonants in some subsequent languages

The situation around AD 500

One of the best-known examples of a sound change that has led from the original language to the individual languages ​​is the development of the three palatal sound series ( palatal , velar and labiovelar , formerly gutturale, now called tectal): In almost all subsequent languages ​​these three tectal groups are closed two coincided. According to the most widespread theory, the palatals were abandoned in the so-called Centum languages (from the Latin centum 'Hundred', also "Labiovelars"); these thus coincided with the simple velars; the labiovelars were preserved. On the other hand, in the satem languages (after avest. Satəm , also "palatal languages") the lip roundness of the labiovelars was omitted; these thus coincided with the simple velars; the palatals were preserved.

In the Centum languages, the labiovelars often developed into labials (e.g. in Celtic and partly in Greek ; sometimes only the rounded lips are preserved, e.g. in nhd. W- and ne. Wh- in the initials of the question words). In the Satem languages, the Palatal often developed into a fricative (e.g. in Ur- Indo-Iranian , in the Slavic languages or in Armenian).

Before the discovery of the Tocharic languages one saw here the after-effects of two Indo-European dialect groups, Kentum in the west (Italian, Celtic, Germanic, Greek) and Satem in the east (Baltic, Slavic, Indo-Iranian, Armenian). However, since both Anatolian and the Tocharian languages ​​localized far to the east are Kentum languages, independent developments in the individual language groups are assumed today. The term Kentum or Satem language has only phonological meaning today.

In addition, the other sounds developed for the original language also experienced more or less strong changes: The voiceless plosives remained largely unchanged in the subsequent speeches, except in Germanic and Armenian, where sound shifts towards fricatives and aspirates took place. The voiced plosives were also only changed in Germanic and Tocharian (they became voiceless).

The voiced aspirated plosives remained only in the Indo-Aryan languages (mostly up to the present day) and in the other languages ​​mostly lost their aspiration or voicing (as in Greek).

Vowels, diphthongs, syllabic resonants and laryngals

Ferdinand de Saussure opened the laryngals in 1878

The five vowels / a /, / e /, / i /, / o /, / u / came in short and long form in Indo-European. (The long / iː / and the long / uː / are not recognized by some, but are attributed to combinations of the corresponding short vowels with laryngals .) The vowels / e / and / o / in short and long form take up by far the largest space here . The resonants / m /, / n /, / r /, / l /, and the laryngals were also used vocalically. The corresponding resonants are then often marked with a small circle under the vowel. Relationships between short and long vowels, consonantic and syllabic resonants and laryngals result morphophonologically from ablaut phenomena .

The diphthongs were / ey /, / oy /, / ay /, / ew /, / ow /, / aw /, and more rarely with long vowel / ēy /, / ōy /, / āy /, / ēw /, / ōw / , / āw /. Instead of the somewhat confusing notation with the half vowels y and w , the full vowel symbols i and u are also used in the diphthong designation (/ ei /, / oi /, / ai /, / eu /, / ou /, / au /); however, this occasionally creates the possibility of confusion with combinations of two full vowels. The semi-vowel notation chosen here makes it clear that the focus of the diphthongs was always on the first component.

Laryngals were onlypreserved directlyin Hittite (there is a and a ḫḫ ). In the other languages, however, there are reflexes in neighboring vowels, most clearly in Greek, where / h₁ / hasno effecton e , but (bychanging the color) / h₂ / an a and / h₃ / an o have caused.


* g̑ʰ á ns 'goose', * m ā -tér 'mother' (note: the word is also reconstructed as * meh₂tḗr), * n ée- l -e h₂ 'cloud, fog', * ph₂t r 'father', * n i -sd-ó- 'nest', * w ey s- (Gen. * wisos) 'poison' ( cf.avest . vīša ), * gʰ o s-ti- 'guest', * wéd ō r 'water', * h₁r u dʰ-ró- 'red', * n ú (± n) 'now, now', * d ey k̑- 'show', * h₁ óy -nos 'one', * k áy kos' blind, one-eyed '(cf. air . caech , got. háihs , Latin caecus ), * t éw -te-h₂' people '(cf. German German ), * l ów ko-' clearing '(cf. Latin lūcus , ahd. lōh ), * t aw r-o-' bull '(cf. Greek ταῦρος taũros ), dative ending * - ōy (cf. Greek - ), athematic nom. * d y-ḗw -s in ved. dyáuḥ , Greek. Ζεύς Zeus , this from the Vok.Sg. * d y-éw Latin Iū- (p) piter , of which as Vr̥ddhi formation * d ey -w-ós 'god of heaven' (Latin deus , dīvus = ved. deváḥ = English Tuesday ), * (d ) k̑ t-ó-m 'hundred', prefix * n̥- (German prefix un -), m -tó- 'dead' (cf. German murder ), m d-ú- 'soft' ( see German mild ).

Development in the follow-up talks

The vowels remained unchanged in Greek for the time being (apart from the mentioned coloring by former laryngals). The u (the Greek Y psilon) became y at the time of Homer or shortly afterwards . In the Ionian and Attic dialect, the long ā became a ɛ: (Greek Eta ). In later developments in Greek the vowel system was greatly simplified by the coincidence of many vowels and diphthongs, mostly to i (cf. Itazism ), whereby the distinction between long and short vowels was also lost. The Italian languages, including Latin, were given the vowels.

In Indo-Iranian, the vowels * e, * o and * a coincide to form a (in the short and long form).

In Germanic the idg. Short vowel * o became a and thus coincided with the old * a; later the Indo-European long vowel * ā was darkened to ō ( ū in final syllables) and in turn coincided with the * ō inherited from the basic language.

Ur-Slavonic - development of the vowels . Short vowels: * e and * o were retained. * a and * ə coincided with o. * i and * u became the half-vowels ь and ъ. Long vowels: * ā and * ī remained as a and i. * ō became a, * ē became ě, * ū became y [ɨ].

Diphthongs in Ur-Slavonic. i-diphthongs: * ai and * oi became ě, in final syllables i can appear as a representative of * oi. * ei became i. u-diphthongs: * au and * ou coincide to u. * eu becomes palatalizing 'u.

Syllabic resonants in Ur- Slavic. * l̩ and * r̩ are retained, * m̩ and * n̩ become nasal ę.

The short diphthongs are continued in Greek, * ow was written to u (but still written as a diphthong ου ), * ey ( epsilon + iota ) to a long e: (also written as a diphthong ει - ei ). The long diphthongs coincided with their initial vowels (in the script the former diphthong character is still recognizable in the iota subscriptum : ). In the development towards Modern Greek, the remaining diphthongs were also monophthongized.

In Vedic Sanskrit the short diphthongs * ey, * oy and * ay first became ai , then a long e , correspondingly from * ew, * ow and * aw over au the long o . (Short e and o are not used). The long diphthongs then became the simple diphthongs ai and au .

The syllabic resonants have lost their syllabic property in most of the subsequent speeches. Scion vowels developed, which sometimes completely replaced the original resonant. So the prefix is * N in Latin to domestic , in the Germanic to un- to and in Greek and Indo-Iranian a- . The syllabic * r̥, also l̩, is still preserved in Indo-Iranian and Slavic (in Vedic also * l̥ with arbitrary distribution from both inherited * r̥ and from inherited * l ebenso, also in Slavonic * l̩ = l̩), but developed later also a scion vowel i (hence the pronunciation Sanskr i t for the language name, in Sanskrit saṃskṛtám 'put together' * se / om-s-kʷr̥-tó-m).


Vedic manuscript with accent symbols (in red)

The word accent is indicated in the Vedas and in Greek in the scriptures. In some other languages ​​(e.g. many Slavic and Baltic languages) the Indo-European accent system has been retained in principle. (However, many individual accents have shifted, systematic accent shifts have taken place, and additional rules have also emerged, such as the restriction to the last three word syllables in Greek.) Nevertheless, the Urindo-European accents often cannot be reconstructed with certainty. It is fairly certain that in the late phase of Indo-European, before the separation into the subsequent languages, the accent was melodic , not dynamic . In addition, it was flexible, which means that the accent position per word was free and not subject to fixed rules (which, for example, resulted from the length of the syllable later in Latin).

The accent position was a difference in meaning: Greek. Φόρος Phoros <* BOR-os, offering ': φορός Phoros <* BOR-ó-s, supporting' or τρόχος trochos <* Drog-os, running, running track ': τροχός trochos <* DRÓG -ó-s 'runner, wheel'.

Many words were enclitic : They did not have their own accent, but rather merged prosodically with the words in front of them. The finite verb forms had the extraordinary peculiarity that they were enclitic in the main clause, but carried the accent in the subordinate clause (so consistently in Vedic).

The accent position had a morphological meaning, especially in the noun, and was used (along with other means such as endings and ablaut) to identify cases.

In Germanic and Italian, the mobile accent was soon replaced by a fixed accent on the initial syllable. Associated with this were phonetic changes in the unstressed vowels, from which one today z. B. can draw conclusions about the original Indo-European accent position ( Verner's law in Germanic). In Latin, the initial stress towards classical Latin was again replaced by the accent rules known today; In Germanic, the initial stress developed into the later principle of stem syllable stress.

Morphology and Morphosyntax

The word

A typical Indo-European word has a structure that is broken down in the traditional description into "root", "suffix" and "ending"; Root and suffix are jointly called stem (the term “suffix” traditionally used here is therefore in a narrower sense than only for derivative suffixes ). In other words, “endings” are inflectional affixes and, accordingly, only apply to inflected types of speech such as nouns, verbs or adjectives.

A comparable formation in German can be found, for example, in read -bar-e (texts) : The root “les”, which also occurs in reading, reading, reading, legible , the suffix “-bar-”, which follows the verb stem and denotes the possibility of performing the respective action, as well as the ending "-e", which stands for the nominative plural.

  • The lexical meaning content is coded in the root ; but it is not limited to a part of speech. Roots are almost always monosyllabic and usually have the structure plosive / resonant / fricative - (± resonant) - vowel - (± resonant) - plosive / resonant / fricative. The “± resonants” may be omitted. Some roots are onomatopoeic, i.e. imitate the action aloud (or vice versa), like * kap or * pak̑ 'grab, snap', * ses 'sleep' or * h₁eh₁ 'breath'. Usually no such connection is discernible; the root is actually a rather accidental sound structure. The coding of the meaning is always firm and clear.

Examples: * h₁es 'be, exist'; * ped verbally 'kick', nominally 'foot'; * gʷem next to * gʷeh₂ 'take a step, come'; * dʰeh₁ 'put, put, put'; * stand₂ 'stand', * deh₃ 'give', * bʰer 'bring', carry '* pekʷ' cook ', * p (y) eh₃' drink ', * melh₂' grind ', * yewg' harness', wekʷ ‚Speak ', * mlewh₂‚ speak', * bʰeh₂ ‚speak ', * leyp‚ stick' or * dʰwer 'door ‚u. v. a. m. Occasionally the initial or final also consist of combinations of consonants without a resonant, e.g. B. * h₂ster 'star' or * sweh₂d 'sweet'.

  • The suffix specifies the meaning in a way that is comparable to the German prefixes (edit, edit). Their semantic function can often no longer be clearly defined, and the suffix often blends with the root and ending beyond recognition.

Examples: * -lo- reduction (cf. Latin - (u) lu-s , - (u) lu-m ), * -ko-, * -iko-, * -isko-: origin, material (Latin bellum “War”, bell ic us “warlike”), Old High German diutisc belonging to the people> vernacular German (as opposed to Latin) .

  • While the suffixes are seen more as elements of word formation, the endings form the main vehicle of the inflection system.

Initially, prefixes only appeared sporadically. The main examples are

  • the negative prefix * n̥-,
  • the reduplication, the preposition of a (mostly shortened) variant of the word root, such as in Latin: present po-sc-ō 'I demand', root po- (in Latin in this sound environment from * pr̥k̑-), in addition perfect po -po-sc-ī, in Greek δί -δω-μι dōmi 'I give'. Reduplication often occurs in conjugation to identify the perfect tense, but also the present tense.
  • the augment , a preceding * h₁é-, which denotes the past in verbs. Since it is only documented in Greek, Armenian and Indo-Iranian, augmentation is assumed to be regionally limited.

In later language levels prefixes came up more often through composition with prepositions and adverbs; they usually remained clearly delimited from the root of the word in the subsequent languages, while the suffixes are mostly fused with the root of the word or the word ending beyond recognition.


The root, suffix and ending of the Indo-European word were subject to ablaut formation. The ablaut system distinguished five levels: the vowelless zero level, the full levels on * -e- and * -o-, and the expansion levels on * -ē- and * -ō-.

Quantitative ablaut
Expansion stage Full level Zero level
e level ē e O
o stage O O

Other vowels were created through secondary formations in connection with these five vowels and laryngals, and above all from the "half vowels" * y and * w, which become * i and * u in the zero stage. Also * m, * n, * l and * r and the laryngals were lengthened in the zero stage to the syllabic sounds with a vowel role. Some elementary * a (e.g. in the roots * albʰ 'white', * kan 'sing', * (h₁) yag̑- 'worship', * bʰag 'allocate', * magʰ 'be able to' or * gʰayd 'goat '), as well as elementary -o-roots such as * pot' mighty ', * gʰos' eat' or * gʷow 'beef', and perhaps elementary * ū in * mūs' mouse ', are known. Other basic vowels than * e in the ablaut with * o are rather rare. The root * swād 'sweet' is certainly preceded by * sweh₂d, as Tocharian shows, where the adjective is in the zero stage * suh₂d-ró- (* -uh₂- urtocharic> * -wa-), i.e. according to phonetic law (* d disappears Consonant) urtoch. * swarë> toch. B swāre 'sweet'; the root * swād is therefore not an example of a basic vowel * ā.

The ablaut was an important element of word formation (Greek λέγ ω légō 'I speak', λόγ ος lógos 'word'), but also of inflection, where it was used to distinguish between person, aspect, case in addition to the accent position and ending.

In the case of a few, all ablaut levels are occupied; such an example is provided by the affinity suffix * - (h₁) ter- in the Greek word for "father":

Quantitative ablaut
Expansion stage Full level Zero level
e level πα τήρ patḗr
Nom. Sg.
πα τέρ .alpha. PATERA
Acc. Sg.
πα τρ ός patrós
Gen Sg.
o stage εὐπά τωρ eupátōr
good as a father,
(epithet of Mithridates VI. )
εὐπά τορ α eupátora
that. in the acc.

Between consonants and the initial sound before the consonant, resonants and laryngals become syllabic in the zero stage, i.e. y > i , w > u , m > , n > , l > , and r > , a laryngeal becomes a Schwa , in usually noted as * ə.

The zero level often results from diphthongs:

  • * trey- 'three': * tri-tó-s 'the third'
  • * k̑weyd- 'white': * k̑wid-ó-s dutch. Niederdt. 'Witt'
  • * g̑ʰew- 'pour': * g̑ʰu-tó-m 'God' (meaning transferred from libation or libation )
  • * dewk- 'draw': * dúk-s Latin dux 'Feldherr' (root noun); * duk-ó-no- (or * -o-nó-) pulled

The so-called 'Full Level II' has the same effect:

  • * (h₁) yag̑- 'worship': * (h₁) ig̑-tó- ved. iṣṭá- 'revered'
  • * swep- 'sleep': * sup-nó-s Greek ὕπνος hýpnos 'sleep'; hittit. šupp (tt) a (ri) 'he sleeps' * sup-ó.

* Em, * en, * el and * er are not referred to as diphthongs despite their structurally identical behavior:

  • * meg̑h₂- 'large': * m̥g̑h₂-éh₂-m Greek ἄγᾱν ágān 'very'
  • * nés 'we': * ń̥s 'us'; likewise * wés 'you': * ús 'you' (root noun; dt. 'you' composed of * us + wés + ge)
  • * g̑ʰel- 'yellow': * g̑ʰl̥-tó-m 'gold' (substantiation by shifting the accent to * -l̥-)
  • * valuable 'contact': * WRT-ó-no (or * -o-NO) Sex been (dt. -d instead -t after the present stem)

There were different developments in the follow-up discussions. All levels are found in Greek, in Vedic * e and * o have merged into a , so that only three quantitative levels remain (known in Sanskrit grammar as the basic level, guṇa and vṛddhi), which occur even more numerous than in Greek . In the Germanic languages, the ablaut in the verbs has developed into the well-known colorful variety of vowels with numerous and especially in German increasingly numerous ablaut patterns (39 in New High German).

According to Rix (1976, pp. 33f.), The origin of the ablaute lies in phonetic effects that have been phonologized and morphologized. Then follows the paradigm formation of the recognized basic rule * -é- under accent, 'zero' under no accent, * -o- if * -é- secondary under no accent (cf. above Akk. Sg. Eupátora , good as father, having a good father '). The difficulty that the "weak" partial stem is therefore difficult to pronounce due to the accumulation of the zero levels is countered by the language by completely omitting initial consonants (Vedic turī́ya- 'the fourth' without initial * kʷ-), forming secondary scion vowels ( * -e- in glbd.Greek τέταρτος tétartos , * -a- in glbd. Latin quārtus ) or uses the means þorn or -n-infix . Such “weak” sub-stems, which have been redesigned secondarily to facilitate pronunciation, can then - in order to achieve a phonologically and morphologically coherent paradigm, cf. Phonologized and morphologized above - “strong” sub-stems are newly formed, which now - but only apparently - contradict the recognized basic rule. So z. B. the "weak" stem of the word 'foot', in the gene. Sg. * Pd-és, pronunciation facilitated by * péd-os and * pod-és, with new “strong” sub-stems (nom. And acc. Sg.) * Pḗd-s / * péd-m̥ (so Latin; Vedic also in Acc. Sg. * pḗd-m̥) or (Nom. and Acc. Sg.) * pṓd-s / * pód-m̥ (so Greek and German.) supplemented to a paradigm that is aural in itself. The parallel development in the verbal paradigm, e.g. B. at the root * h₁ed 'eat', confirms this phonological-morphological origin hypothesis: * h₁d- in tooth (* h₁d-ónt-), * h₁ḗd-ti / h₁éd- (o) nti ultimately in Latin ēst and edunt , * h₁ṓd-mi / h₁od- (é) nti in Armenian owtem 'ich esse' and in the German causative I ätz * h₁od-é-ye-.

It is similar in German (and to a lesser extent in English) with the effect of the umlauts, which is independent of the ablaut ( man - men , man - men , I run , you run ), which arose from a vowel harmony effect and later to differentiate between grammatical forms served.

Theme vocal

A common suffix, but without a tangible meaning, is the so-called theme vowel * -e - / * - o-. If it occurs between stem and ending, the corresponding inflection paradigms are called “thematic”, otherwise “athematic”. The athematic inflections are more complicated than the thematic, mainly because of the phonetic effects between stem and ending. In the course of time, more and more verbs passed from the athematic to the thematic classes in the subsequent speeches. In the noun, the thematic class in Latin and Greek is the o-declension. The athematic verbs in Greek are the "Verba on -μι (-mi)" (for example: δίδω-μι dídō-mi 'I give' <* dé-doh₃-mi), in Latin a few irregular verbs like esse 'sein ', velle ' want 'or īre ' go '. The so-called "consonant" or "3. Conjugation “of Latin (e.g. dīcere 'say' * déyk̑-o-) is not athematic, but a short vowel e-conjugation in contrast to the long vowel ē conjugation (e.g. monēre 'mahnen' * mon-é -yo-; see the following text), and the direct continuation of the Indo-European thematic conjugation.

According to the synchronous reconstruction findings, the theme vowel * -e - / * - o- does not adhere to the ablaut rules and is also immune to fading in unstressed syllables. Rasmussen's suggestion that the theme vowel * -o always occurs when the following sounds are voiced, for example * bʰér-e-si 'you carry', * bʰér-e-ti 'he carries', but * bʰér- o-mes 'we carry' and * bʰér-o-nti 'they carry', too many contrary cases stand in the way of realizing the theme vowel, e.g. B. the nom. Sg. Of the thematic stems on * -os (not * -oz) or the pronominal genitive about Latin cuius from * kʷó-syo (not * kʷó-zyo). The change between * e and * o cannot actually be traced back directly to ablaut . Rix (1976), however, rightly considers deriving the theme vowel in the noun from the athematic ending of the gene. Sg. * -És (interpreted as * -é-s and transferred to the other paradigmatic forms) and in the verb from the athematic ending of the 3rd person plural * -énti (interpreted in the same way as * -é-nti and to the transferred to other paradigmatic forms). Acrostatic and proterokinetic inflection types (for the terms see below accent and ablaut classes ) regularly generated the ablaut correct endings Gen. Sg. * -Os and 3rd pl. * -Onti, interpreted as theme vowel variants * -os and * -o-nti. Both variants were then distributed in the paradigm according to the reconstructive findings.

The long vowel conjugation classes in Latin have different origins. The Latin ē conjugation ("2nd conjugation") consists of root verbs (e.g. -plēre 'fill', nēre 'spinnen'), causative iteratives on * -é-ye- (e.g. monēre ' to warn ' * mon-é-ye- , see old Indian mānáyati , or docēre' to teach '), tripod verbs on * -éh₁-ye- (e.g. sedēre ' sit ', vidēre ' see '), denominatives on * - é-Ye / * -e-Ye (z. B. fatērī , acknowledge ', Salvere , be healthy'), and on denominatives * it-Ye (z. B. decēre , befit 'to decus , Zierde 'or augēre ' multiply 'to * h₂éwg-os in Latin augus-tus ' sublime' and ancient Indian óyas ' power ').

Also some verbs of the ā-conjugation ("1st conjugation") have their origin in causative-iterative formations, where (root-ending) * h₂ the following * e changes color to * a, z. B. tonāre 'thunder' * tonh₂-á-ye- <* tonh₂-é-ye- or domāre 'tame' * domh₂-á-ye- <* domh₂-é-ye-. There are also inherited factitives from adjectives to * -éh₂- (ye-) / * -eh₂- (ye-), z. B. novāre 'renew' or aequāre 'equalize'. The main source of ā-conjugation, the pattern of which was decisive for the development of the regular paradigm of ā-conjugation, are denominatives on * -éh₂-ye- / * -eh₂-ye-, which can be derived from the formations of the collectives. The ā-conjugation also has root verbs such as nāre 'swim', flāre 'blow' or fārī 'speak'.

The verbal suffixes * -yé - / - yó- or * -ye - / - yo-, which form the 4th verbal class in ancient Indian and which are also widely used in other Indo-European languages, lead to the development of the ī conjugation (“4th . Conjugation ”), but only after the“ heavy ”syllable ( audīre ' to hear'); after the "lighter" syllable, the corresponding verbs are added to the "3. Conjugation "incorporated ( capere 'seize'). The ī conjugation also includes further training with special semantics, e.g. B. the suffix -urīre , which always expresses the intention of an action: ēsurīre "want to eat" d. H. “To resemble someone who eats” (to edō ' to eat'). These patterns were generalized as Latin evolved.


Nouns were inflected according to number and case and classified according to gender .


There were three numbers: singular, dual, and plural. The dual denotes a twofold number of objects. It is assumed (mainly because of the absence of the dual in Hittite) that the dual did not yet exist in earlier language stages, and then via the designation of natural pairs (for example, paired body parts) and the dialogue situation based on the two people you and me originated. In the following conversations the dual died out almost everywhere; Obviously, the longest it stayed in the inflection of words like “two” or “both”. In Vedic one sees the dual as a number for the general designation of the two number, in Greek it was only used for natural couples. The old Germanic languages ​​such as Gothic , Old Norse or Old High German still retain the dual. In Gothic it is even present - albeit to a limited extent - in verbal inflection. The Old Norse personal pronouns nominative vit , genitive okkar , dative / accusative okkr "we both" and N. þit , G. ykkar , D. / A. ykkr "you both" have corresponding counterparts and the like. a. in the West Saxon dialect of Old English ( wit - uncer - unc - unc ; ȝit - incer - inc - inc ) and also Gothic ( wit - * ugkara - ugkis - ugkis ; * jut - igqara - igqis - igqis ). In Old High German, on the other hand, only the genitive of the first person, unrecognized , is used once. Formally, however, the series of forms of the 2nd person dual lives on in some German dialects ( Bavarian and South Westphalian ), albeit with a plural meaning (example: bair. Ees - enker - enk - enk ). Even the Icelandic received the dual-form series. However, there was also a reinterpretation towards plural pronouns.

For the early Indo-European original language, another number is assumed to denote collectives, i.e. to denote a multiplicity of objects as a unit (for example "humanity" as opposed to "people"). The two plural forms locī (e.g. 'places in books') and loca ('area') of locus ('place') can be found as relics in Latin , or in Greek κύκλοι kýkloi ('individual wheels') and κύκλα kýkla ('Gear train, set of wheels') from κύκλος kýklos ('wheel'), where loca and κύκλα kýkla each denote the collective.


Based on the eight cases of the Vedic, one also assumes eight cases for the Urindo-European. These are the nominative (subject of the sentence), the vocative (addressed or called person), the accusative (direct object of the sentence, goal of the movement), the instrumental (means, tool), the dative (indirect object, beneficiary), the Ablative (starting point of movement, reason), the genitive (nominal attribute, affiliation, area), and the locative (place of the object, indication of time). A possible ninth case, the directive or allative (goal of the movement), is discussed in view of some traces in Old Ethite.

The number of cases has decreased in the subsequent interviews. In Latin, for example, the instrumental, the locative (apart from a few traces) and the ablative coincide to form a single case "ablative". In Slavonic there are still seven cases, here the ablative is fused with the genitive.

The two Tocharian languages ​​form a special case, in which the number of cases has even increased. However, only four of the cases go back to Indo-European; the others are innovations triggered by agglutinating neighboring languages.


There were three genera in Indo-European, masculine, feminine and neuter. Based on the Hittite findings, it is assumed that the division into masculine and feminine did not exist in the early phase. Instead there were animata and inanimata , that is, animate subjects and inanimate objects. The Inanimata became the Neutra, while the division of the Animata into masculine and feminine, presumably in connection with a division into male and female sex, was formed over time.

The Inanimata (neuter) could not be the subject of a sentence, so there was no nominative for them. This can still be observed in the following conversations with the neuter, where the accusative (or in Hittite a case that goes back to the instrumental) takes on the role of the nominative.

It is believed that Inanimata only had the collective plural. One trace of this would be the Greek phenomenon that when a subject is in the neuter plural, the verb is in the singular.


Ending schemes

The following table shows reconstructed ending schemes including the characteristic suffixes.

Thematically Athematic
o stems Cons. eh₂ strains i tribes u strains
m (f) n m / f n f m / f n m / f n
Nominative -o s -o m - s , - ø - ø -eh₂- ø -i- s - ø -u s - ø
vocative -e- ø - ø -eh₂- ø -ey- ø -ew- ø
accusative -o m - -eh₂- m -i- m -u m
Instrumental -o h₁ , -e- h₁ - (e) h₁ -eh₂- eh₁ -i- h₁ -u h₁
dative -ōy (<-o- ey ) - ey -eh₂- ey -ey- ey -ew- ey
ablative t - s , -e s , -o s -eh₂- es , -eh₂- os -oy- s -ow- s
Genitive -o- s (y) o
locative -o- y , -e- y - i , - ø -eh₂- i -ēy- ø -ēw- ø
Thematically Athematic
o stems Cons. eh₂ strains i tribes u strains
m (f) n m / f n f m / f n m / f n
Nominative -ōs (<-o- es ) -e- h₂ - it - h̥₂ -eh₂- it -ey- it -i- h₂ -ew- it -u h₂
accusative -o- ms - m̥s -eh₂- ms -i- ms -u ms
Instrumental -ō- ys , -o- mis - bis , - mis -eh₂- bʰis , -eh₂- mis -i- bʰis , -i- mis -u up , -u mis
dative -o Bos , -o mos - bʰos , - mos -eh₂- bʰos , -eh₂- mos -i- Bos , i- mos -u Bos , -u mos
Genitive -ōm (<o- om ) - om -eh₂- om -y- om -w- om
locative -oysu - see below -eh₂- see below -i- su -u su

Hardly any statement can be made about the dual, except that the ending in the nominative / vocative / accusative * -h₁ or * -e should have been.

The * -i and * -u stems behave like other athematic nouns and do not actually form their own declension classes. In many subsequent conversations, however, they have developed a life of their own through fusions and analogies.

With the thematic (* -o-) declination, the ending sentences have moved more and more away from the athematic endings over time. The genitive in Latin and Celtic is conspicuous , which led to the (now rejected) assumption of an Italo-Celtic subgroup of the Indo-European languages.

The (athematic) * -eh₂-feminines are the origin of the ā -declinations of the various subsequent languages ​​(in Vedic the thematic * o-declension has become the not to be confused with a -declination, the feminine ends in long ā ). Since these stems often form the female versions of the masculine words of the * o-stems, the ending schemes were adjusted in the subsequent languages. A variant of the * -eh₂-feminine are the * -yeh₂-feminine, which led to the large group of -ī́ feminine (e.g. devī́ 'goddess') in Vedic.

The Latin masculine job titles on - a ( poēta 'poet', agricola 'farmer', nauta 'sailor', scrība 'writer') consistently follow the declination scheme of the ā -declination, just as many male personal names (e.g. Sulla , Cinna , Catilīna or Caligula ). In Greek, however, such job titles and personal names are given an -s in the nom. Sg. And take over in the gene. Sg. The ending -ou from the -o declination, e.g. B. οἰκέτης oikétēs 'servant', πολίτης polítēs 'citizen', δικαστής dikastḗs 'judge' or Αἰνείᾱς Aineíās , Λεωνίδᾱς Leōnídās or Ἀτρείδης Atreídēsς .

Accent and Ablaut classes

In addition to the endings, the case of the athematic nouns is marked by the position of the accent and the ablaut level of the elements root, suffix and ending. This older system is still clear in Vedic and Greek, in Latin it is still rudimentary in the difference between the nominative stem and the stem of the other cases in the consonant declension (for example Latin noun , nominis ).

A distinction is made between strong cases and weak cases . The strong cases are nominative, vocative and accusative in the singular and dual, nominative and vocative in the plural; all other cases are weak. The locative singular is mostly (and probably originally) strong; but it can also appear as a weak case. The four main classes of declination are called acrostatic , proterokinetic , hysterokinetic and amphikinetic . Instead of -kinetic , the term -dynamic is also used; there are also other declination classes such as B. mesostatic . The following table shows typical examples.

  acrostatic proterokinetic
root suffix Ending root suffix Ending
strong cases Accent
* é level
* h₂éw-is (Latin avis 'bird')
* é level
* péh₂-wr̥ (Heth. Paḫḫur 'fire')
weak case Accent
* é level
* h₂éw-is (Gen. Latin avis 'des Vogel')
* é-level
ph₂-wén-s (Gen. heth. Paḫḫuenaš 'of fire')
hysterokinetic amphikinetic
root suffix Ending root suffix Ending
strong cases Accent
* é-level (Acc.Sg .; Nom.Sg. with * ḗ-level)
* ph₂-tér-m '(' father ')
* é level
* h₂éws-ōs ('Dawn')
* ō-level (Nom.Sg.)
* o-Level (Acc.Sg.)
weak case Accent
* é level
* ph₂-tr-és (Gen.)
* é level
* h₂us-s-és (Gen.)

The empty fields indicate the unstressed zero level.

A particularly common declination class is the mesostatic (accent in both the "strong" and the "weak" sub-stem on the suffix), z. B. ai. matíḥ , Gen. matéḥ ‚thought '* mn̥-tí-s, Gen. mn̥-téy-s, or ai. víḥ , Gen. véḥ 'bird' * h₂w-í-s, Gen. * h₂w-éy-s (cf. also in the following section the examples for root nouns ). The declination class is considered productive education and is therefore only archaic in individual cases.

In the so-called root nouns, the root is in the zero stage, but has the accent (e.g. Latin nix 'snow' * snígʷʰ-s, German castle * bʰŕ̥g̑ʰ-s). They often have an equally standing in the zero grade suffix (eg. As lat. Portus * PR-tú-s, harbor 'with -tu suffix, dt. Overall burt * BR-tí-s with Ti suffix) , which then usually bears the accent. They come with (meso-) static (if the suffix in the “weak” sub-stem is in the accented -é- or -ó-stage, e.g. Gen.Sg. 'des Hafens' * pr̥-téw-s, der overall burt * BR-toy-s) and mobile (= amphikinetischen if the ending is emphasized in the "weak" part master, for. instance gen, the snow '* Snig-és, the castle * BRG és ) Accent.


Like nouns, adjectives were declined according to number and case, but unlike these, also according to gender. The forms are the same as those of the nouns (and an adjective could also be used as a noun). Most of the adjectives in masculine and neuter follow the * o-declension, in feminine the * eh₂- or * yeh₂-declension. * I- and * u- or consonantic adjectives were also used; the feminine was sometimes designated by the * yeh₂ suffix, sometimes not at all.

Adjectives can also be increased . The comparative is denoted by the amphikinetic suffix * -yos (ablaut forms * -yōs * -is) (Latin. Maior , 'greater') or the thematic suffix * -tero- (Greek: σοφώτερος sophṓteros , 'wiser'). The superlative has the suffixes * - (m̥) mo- (Latin. Minimus 'the smallest') or * -is-to- (Greek μέγιστος mégistos 'the greatest'). The Latin. The superlative ending -issimus is based on a combination of the comparative suffix * -is- with the superlative suffix * - (m̥) mo-.


The reconstruction of the various forms of pronouns is only partially possible.

Personal pronouns

The personal pronouns of the first and second person (for the third person see under demonstrative pronouns ) had no gender distinction. There were singular, dual, and plural; However, it must be noted that “we” is not the plural of “I” in exactly the same sense as “persons” is the plural of “person”, since the roles of the speaker and the person addressed cannot easily be included in these categories . Correspondingly, there are completely different roots in the singular than in the plural.

The personal pronouns each had a stressed and an enclitic form. This distinction has persisted in Greek and Indo-Iranian; In other subsequent talks the forms of the two types have mixed up. The enclitic form did not appear in all cases.

  I you we her
stressed encl. stressed encl. stressed encl. stressed encl.
Nominative eg̑óh₂, eg̑h₂óm - tú (h₂) - wéys - yúhs -
accusative mḗ me tḗ te, twe n̥smé nos usmé what
dative meg̑ʰey, meg̑ʰyom moy tebʰey tebʰyom toy n̥sméy ? usméy ?
Genitive méne téwe n̥sóm usóm

reflexive pronouns

The dative form * soj and the enclitic accusative * se can be reconstructed from the reflexive pronoun * swe / * se.

Demonstrative pronouns

As in modern languages, there were various demonstrative pronouns that expressed different types or degrees of demonstrativity. (Compare he, this, that, the same).

The pronoun * só / * séh₂ / * tó- ('he' / 'she' / 'es') became - in attributive use - the starting point of the definite article in Greek, in a certain sense also in Vedic, much later also in German . The forms shown here are mainly reconstructed based on the Vedic inventory:

  Singular Plural
m n f m n f
Nom. so death sé-h₂, sí-h₂ tó-y té-h₂ té-h₂-s
Acc. tó-m té-h₂-m tó-ms té-h₂-ms
Date tó-smo-ey tó-sye-h₂-ey tó-y-bʰyos té-h₂-bʰyos
Oj tó-smo-et tó-sye-h₂-s
Gene. tó-syo tó-y-sōm té-h₂-sōm
Locomotive. tó-smin tó-sye-h₂-m tó-y-su té-h₂-su

This pronoun can be found in the German das , in the Greek article ὁ ho , , τó tó , and in the Vedic pronoun , sā́ , tád .

A second demonstrative pronoun * i- (ablaut * ey-) corresponds to the Latin is , ea , id , Sanskrit ayám , iyám , idám

Interrogative, indefinite, relative pronouns

As interrogative pronoun is a noun * AI (AI * s lat. Nom. Mask. Fem. Quis ? Who '= Greek. Τίς TIS , * AI d lat. Nom. Acc. Neuter. Quid ? What?' = Greek τί ) and adjectival * kʷó- reconstructed. From the fact that the question pronoun has its own form for the neuter, but no gender distinction between masculine and feminine, one concludes that these forms are very old. However, the Vedic knows the stem * kʷí- only in one form, namely the nom. Akk. Neutr . Kím ? ,What?'; the other forms (which then have an alternative neuter in kád ? = kím ? 'what?') go exactly like the demonstrative pronoun tá-.

In enclitic form, the question pronouns had an indefinite meaning (“whoever”).

The relative pronoun also goes back to the question pronoun and develops z. T. own forms. Another relative stem was * yo-, possibly with an introductory laryngeal (h₁); This is known in Sanskrit as the relative pronoun yád , in Greek as , in Celtic as yo .

Interrogative pronouns and relative pronouns can be comprehensively represented reconstructively by replacing the demonstrative pronoun * t- with * kʷ- for the interrogative pronoun in the table above and * t- with * (h₁) y- for the relative pronoun. In the interrogative pronoun there are, however, the apparently older "additional forms" mentioned above from the pronominal stem * kʷí-. In this context, the instrumental formation * kʷí-h₁, which is still in Latin quī ? ,how?' and engl. why ? ,Why?' is preserved.

  Singular Plural
m / f n m / f n
Nom. kʷí-s kʷí-d kʷéy-es kʷí-h₂
Acc. kʷí-m  
Instr. kʷí-h₁  
Date kʷó-smo-ey kʷé-smo-ey    
Gene. kʷó-syo kʷé-syo    

More pronomial formations

Adjectives corresponding to the possessive pronoun were reconstructed. The genitive of the personal or demonstrative pronoun usually takes on this function. Further words (another, none, the numerals etc.) fit into the system of pronouns in terms of their role and inflection.


The Indo-European verb was inflected according to number, person, aspect , tense / mode and diathesis . In addition, there were more or less productive procedures that (mostly with a suitable suffix ) enabled the formation of new derived verbs (for example causative , desiderative ). Other suffixes allowed the formation of verbs from nouns / adjectives ( denominative ) or, conversely, the formation of adjectives / nouns from verbs ( participle , gerundive , gerund , etc.).

It is assumed that in a preliminary form of Indo-European the suffixes for tense, aspect, action type etc. could be combined more freely, so that it is not possible to separate between word formation and inflection. From this the “classical” Indo-European verbal system developed, which can be found in its full expression especially in Greek and Indo-Iranian. In some subsequent languages ​​(for example Latin, more distantly Germanic) one can see a later reorganization of this system, in the case of Hittite one rather assumes that the classical system only developed after the language split off.

Number and person correspond to what is known from modern Indo-European languages, with the number of the dual being added, of course.


The most important category of the Indo-European verb is not the tense (as the term “verb” for “verb” might suggest), but the aspect. The aspect expresses the speaker's attitude towards the reported process: the perfective aspect sees the entire course of action in its classification in the report process ("completed action"), in the imperfective aspect the reported point in time lies within the course of action, and in the resultant aspect is the Report focused on the outcome of the process.

The three aspects correspond to the Indo-European form groups present (imperfect), aorist (perfective), and perfect (resultative); (The term "Tempus" should be avoided here). The perfect, however, has a special position due to its genesis (see below the explanation of the factive endings of the past perfect as well as the perfect medium representation in the section The verb in the subsequent languages: Greek ). For the development of the specific semantic characteristics of the aspects cf. the explanations in the following paragraph tense / mode .

It is believed that there were two types of verbs (or actually two different parts of speech) in an earlier language level: the factive verbs and the tripod verbs. The factiv verbs denote one-off events and actions, the tripod verbs long-term states. The factiv verbs are transitive, the tripod verbs intransitive. There is speculation that the factiv verbs are connected with the animated nouns, the tripod verbs with the inanimate nouns. The factiv verbs have the forms and endings of the later present and aorists (without medium), the tripod verbs have the forms and endings of the later tripod and medium. The later perfect has the form stock of the factive verbs and the ending stock of the tripod verbs.

Deviations from the assignment of factual verbs: factual endings and tripod verbs: tripod endings are numerically rare; but they concern important single verbs, tenses and verbal groups, e.g. B. on the one hand (tripod verbs with factual endings) lat. Est 'is', it 'goes' or fit 'will' (in Lat. But semideponens) or the Greek Passivaoriste, on the other hand (factive verbs with tripod endings) transitive depository such as ved. sácate = Greek. ἕπεται hépetai = Latin. sequitur 'follows', or the singular of the hittite. ḫi verbs. These are semideponents with a unique, but in this individual language regularly grammaticalized distribution of factual and tripod endings. This grammaticalization leads to the astonishing phenomenon that the Hittit. Faktitiva Hi verbs are.

In a visionary way, Pedersen identified the factive verbs or the stand verbs in their properties that are eminently important for the knowledge of the verbal relationships in the Indo-European original language as early as 1933 and gave them the term mi-conjugation or H-conjugation . A more general, but meaningful name is urractive or original medium .

Morphologically, the aspect is expressed by the formation of separate stems for present, aorist and perfect from the root of the word. The formation methods are various combinations of ablaut levels, reduplication and special suffixes. The perfect is also characterized by a separate set of endings.

Tense / mode

Within an aspect group (in the perfect tense but not fully developed) there are five tense / mode categories: the present (missing in the aorist group for logical reasons, since a present process is not yet complete), the past, the subjunctive (which represents the future or the intention), the optative (desire, possibility), the imperative (command, not in the first person). Served for identification

  • the ending clauses:
    • the so-called primary or hīc-et-nunc -endings for present and subjunctive,
    • the secondary endings for past and optative,
    • a special ending sentence for the imperative;
  • the augment to mark the past (is regarded as a purely Greek-Armenian-Indo-Iranian peculiarity);
  • special suffixes:
    • * -e - / * - o- or .. * -é - / * - ó- (the subject vowel) for the subjunctive,
    • * -yéh₁ - / * - ih₁- for the optative.

In place of the “aorist of the present”, which is not possible in terms of language typology , there is the injective , i.e. H. an augmentless aorist who represents a non-temporal consideration of the aoristic state of affairs, d. H. an action with “duration zero” (corresponds to the perfect aspect ). This action cannot actually be represented verbatim because the statement requires a longer period of time than the action (e.g. the balloon bursts ). The respective verbal form cannot express the present; accordingly it cannot have any primary endings. Real linguistic representations of the aoristic facts are possible in the past tense, in the future tense and in the modes ( the balloon burst / burst / burst / will burst / can burst / if it bursts we are frightened ). The iteration is also linguistically real ( several balloons burst one after the other ; now with an imperfect aspect ); To express this iteration, aorist roots often form present tense stems with iconic reduplication. The iterated form then corresponds to the iterated semantics; the primary ending can take place without any problems (cf. here in the article “Vedic and Sanskrit” ved. jáṅ + gan + ti 'he comes' * gʷém + gʷom + ti; the aorist root * gʷem 'can take a step' through iteration as the original To express meaning: 'he takes steps').

Aoristic actions that have taken place can, in retrospect, only ever be viewed as a total action because of their “zero duration” . From this develops the meaning of the aorist as an overall view of long-lasting past actions. The present tense then semantically denotes the course of events , i.e. H. Actions that are in progress (often while other actions are occurring). The term aspect relates to the distinction between “overall view ” and “progress view ”.


We know the diathesis active-passive from the modern Indo-European languages , which has developed independently in the individual language branches. A formally separate passive voice did not exist in the original language; Instead there was a medium that denoted intransiveness (lat. abdor , 'I am ...' or '... lie hidden'), as well as reciprocity (lat. abduntur , 'they hide each other'), and furthermore that the subject of the sentence is also a direct or indirect object (Latin abdor 'I hide myself' or 'I hide myself' - the latter meaning, that of interest, is no longer recognizable in Latin, however). From this typical media content, meanings such as (gerundival) Latin abdor 'I let myself be hidden' or (passive) Latin abdor 'I will be hidden' could develop.

Ending schemes

When trying to develop the formal nature of the basic language verb endings, it can be assumed that - in the individual languages ​​e.g. Sometimes in completely different ways - chronological gradations, changes and further developments make it practically impossible to represent the verb endings in a single table. Nevertheless, some facts about the traditional "ending material" are fairly certain:

1. The original ending clauses are largely known; they are - with the restriction largely - for the factual endings 1.Sg * m, 2.Sg. * s, 3rd sg. * t, 1st du * wé, 1st place * mé, 2nd place * té, 3rd place * ént, for the tripod endings 1.Sg * h₂e, 2.Sg. * th₂e, 3rd sg. * e, 1st du * wé, 1st place * mé, 2nd place * é, 3rd place * ŕ̥.

2. So-called hīc-et-nunc markings turn these “secondary endings ” into “primary endings ” (for how to use them, see, for example, aspect and tense / mode ). These markings are obviously * i, * s, * h₂ and u. U. also * r (if this penetrates other tripod endings). The generally and widely recognized distribution of the marking leads tentatively to the following primary endings: factive endings 1.Sg * mi, 2.Sg. * si, 3.Sg. * ti, 1st you * wés, 1st place * més, 2nd place * th₂é, 3rd place * énti, tripod endings 1.Sg * h₂ey, 2.Sg. * th₂ey, 3rd Sg. * ey, 1st you * wés, 1st place * més, 2nd place * éy, 3rd place * ŕ̥s.

3. The endings are partly resistant to ablaut; when they ablaod, they are often used independently and without reference to the valid accent-ablaut assignment.

4. The 3rd place * ŕ̥ of the tripod ending sentence takes a striking development: In the Greek. and German. it is completely eradicated; in the Vedic it penetrates z. T. in the 3rd place of the root aorist. In the Hittite. it appears in all 3rd-pl. forms of the past tense and is partially (and variably) transferred to other endings of the medium and passive. In Latin this transmission takes place almost completely, in Tocharian it takes place continuously. It can be assumed that such transmissions are secondary to individual languages. In the following table of endings, * -ŕ̥ is only used in the 3rd place. of the medium.

5. The tripod endings suffer their greatest changes or losses as a result of the fact that they are transferred from the originally uniform tripod into the medium and (later) into the perfect (with the perfect still retaining the original form of the tripod end). E.g. the 3rd Sg. * e in the medium as * ó ( sīc !) and under the influence of the factual ending 3.Sg. * t as * tó (the 2.Sg. then accordingly as * só instead of * th₂é, the 3rd place as * -ń̥to * -stead of * -ŕ̥ etc.). This 3rd Sg.-ending * tó is additionally provided with * r in Latin (see above) and thus receives the familiar form * -tur . Heth. šupp (tt) a (ri) 'he sleeps' shows this process in parallel in the same individual language: šuppa * sup-ó, šupptta * sup-tó, šuppttari with * r plus hīc-et-nunc marking.

6. Endings of “strong” sub-stems are not accented because the typology of the “strong” sub-stem already bears the accent. This applies to the factual endings of the aorist and present tense and the tripod endings of the -é-tripod and the perfect tense (and without exception, of course, also for the nominal endings in the nominal system). The medium of the aorist and present tense arises from the fact that the originally uniform stand gives its “weak” part of the trunk into the aorist and the present tense (see above under 5.). The now necessary filling of the medial paradigm leads to the unique phenomenon that strong stem endings are accentuated in the medium (so now 1.Sg. * h₂é, 2.Sg. * th₂é, 3.Sg. * é; see e.g. below in the section The verb and the subsequent conversations , Greek the medial form e.g. 1.Sg. * dʰe-dʰh₁-h₂éy 'am set' or 'have been set'). The same phenomenon naturally occurs with the zero-stage tripod, i.e. H. in those cases in which the tripod does not carry out the -é stage but the zero stage uniformly in the entire paradigm.

7. The 1st Sg. of the thematic verbs on * -ō. In the traditional derivation from * -o-h₂ (h₂ but not sure) it is noticeable that this common and important form has no hīc-et-nunc mark. Furthermore, for all securely reconstructed endings without exception, the secondary ending is the primary ending minus the hīc-et-nunc marking (e.g. * si - * i = * s). The secondary ending of the 1st Sg. but also the thematic verb is * m. It appears possible that * -ō is a continuant from * -o-mh₂. The thematic 1.Sg. would then have the hīc-et-nunc particles * h₂, and it would apply * mh₂ - * h₂ = * m. In the monosyllabic, the continuant could actually be * -ó-m (Latin sum 'I am' then from * h₁s-ó-mh₂ instead of ** sō ; Latin 'I give' is not a “real” monosyllabic, but a present tense reduplication , see re-d-dō 'I give back').

8. An "-s addition" for the tripod endings 1.Pl. * -mé + dʰh₂ and 2nd place * -dʰw + é (so in Tichy-2000, p. 86) is not accepted. The only Greek ending of the 2nd place. of the medium, -σϑε , is abstracted from the Narten form of the root * h₁es 'to be', ἧσϑε hḗsthe 'you sit' * h₁ḗs-dʰwe, interpreted as * h₁ḗ-sdʰwe (in Greek is then in the whole paradigm only the stem * h₁ḗ- carried out, with the exception of the significant double form 3rd Sg . Imperfect ἧστο hḗsto and ἧτο hḗto .)

9. Many individual language endings are formally different; one tries, however, to comprehend the development history conclusively. The 1st place the factive primary endings are Greek -μεν -men ; here * n replaces the hīc-et-nunc -particles * s. In the Hittite. it reads -u (u) eni (with the same hīc-et-nunc marking as in Greek); here is * w from the 1st you. transferred and the hīc-et-nunc -particles * i added again. The 1st Sg. the tripod secondary endings are Greek -μην / -μᾱν mēn / mān ; the composition can be imagined as follows: * m + * h₂a + laryngeal (which one?) + mentioned (here meaningless) hīc-et-nunc -particles * n. In the 2nd place of the tripod endings in Latin. replaces an infinitive ending (infinitive in the sense of a request) -minī the inherited ending etc. etc.

10. The parts of the secondary endings highlighted in the table confirm the connection with the perfect endings.

Under these restrictions and very difficult conditions, an extension table could look like this:

  Active ("factual endings") Medium ("tripod endings")
number Pers. Primary Secondary Primary Secondary
Singular 1. -mh₂ (them.), -mi (athem.) -m -h₂ey -h₂e
2. -si -s -th₂ey -th₂e
3. -ti -t -ey -e
dual 1. -wés -wé -wé + dʰh₂ +? -wé + dʰh₂
2. (-th₂és) (-téh₂) (-th₂éyh₁) (-th₂éh₁)
3. (-tés) (-téh₂m) (-téyh₁) (-téh₁)
Plural 1. -més -mé -mé + dʰh₂ +? -mé + dʰh₂
2. -th₂é -té -dʰw + éy -dʰw + é
3. -énti -ént -ŕ̥s -ŕ̥

The endings in parentheses must be considered fairly speculative.

For the imperative, only the singular endings in active can be reliably reconstructed. Ending of the second person singular imperative is 'Null' for thematic verbs, * -dʰí for athematic verbs ('Null' also occurs occasionally, however, 'athematic', e.g. Latin ī ! 'Go!' * H₁éy: * h₁i- dʰí in ved. ihí , altavest. idī , Greek ἴϑι íthi or hethit. īt ). In the 2nd Sg. and 3.Sg. there is the ending * -tōd for requests in the future, e.g. B. Latin petitō ! 'You should demand', 'he should demand' * pét-e-tōd. For the forms of the other persons, numbers and diatheses, the corresponding injective formations were used.

The perfect endings are (see above) identical in origin to those of the medium, but have an active function (due to the history of the perfect tense). Tichy-2000, pp. 89f. also takes primary endings 1.Sg * -h₂ey, 2.Sg. * -th₂ey, 3rd Sg. * -ey, 1st place * -més and 3rd place -ŕ̥s on. The following endings are then secondary endings; they can be reconstructed with a very high degree of certainty as follows:

1. Sg. -h₂e
2nd Sg. -th₂e
3rd Sg. -e
1. You. -wé
1st pl. -mé
2nd pl.
3rd pl. -ŕ̥

The past of the perfect, the past perfect , has the F active secondary endings * -m * -s, -t *, etc. Assuming that the perfect tense of the (reduplicated) Present originated by replacing the stand endings of Faktivendungen introduced the past perfect actually still shows the original ending inventory.

There are no infinitives in the basic language; The individual languages ​​use nominal suffixes to form their infinitives, usually with the case forms of the accusative, dative, locative, etc.


In Greek, Indo-Iranian, Phrygian and z. T. Armenian (see also under Balkan Indo-European ) appears in the past tenses as a marker for the past, a special prefix, the so-called augment; see. Greek é-phere = ved. á-bharat 'he wore' ( past tense ) or in the Armenian aorist form e-ber 'he wore' (in the 1st person singular but beri without augment). In the other common languages, such as Latin or Germanic, the augment is missing. In addition, augmentation was not compulsory in older Indo-Iranian or in Homeric Greek (these non-augmented past tenses are known as injectives).

Meier-Brügger used an adverb * (h₁) é for Urindo-European, which could optionally appear in front of the corresponding verb forms in the past. The above-mentioned Greek ( é-phere ) and Vedic example ( á-bharat ) is reconstructed in Meier-Brügger as * h₁é * bʰéret, contracted as * h₁é-bʰeret.

Stem formations


The formations for present tense stems in Indo-European are manifold. Therefore only the most important are mentioned here:

  • * -yé - / - yó- or * -ye - / - yo-: This suffix, which results in a thematic stem, can probably be considered the most productive in Indo-European. The root is either in the zero level, if the verbs are mostly intransitive, or in the full level, which usually results in transitive. The suffix is ​​also often used to form denominatives .
  • * -é-ye - / - é-yo-: These two suffixes may be variants of the above. The root tends to be in the o-stage and the meaning is either causative or iterative .
  • * -sk̑é - / - sk̑ó-: This thematic suffix is ​​attached to the zero level of the root and gives stems of iterative meaning. For example, go Inchoativa of Latin, with -sc are aktionsartspezifiziert, in this form back, as are the iteratives with -sk̑é * - / - sk̑ó- in Greek and Hittite.
  • * -h₁s (y) é - / - h₁s (y) ó- or * -h₁s (y) e - / - h₁s (y) o-: This suffix occurs either at the reduplicated root (for example * dʰedʰh₁- of * dʰeh₁-) or to the * e-level and has desiderative meaning. It is the origin of some Indo-European future formations, as grammaticalized in Greek.
  • “Nasalpräsens”: An infix * -né- (in the “strong” sub-stem), starting with * -n- (in the “weak” sub-stem), was inserted in front of the last consonant in the zero level of the root. The resulting trunk was originally athematic, but was thematized in a variety of ways in the subsequent discussions. The nasal present is u. a. still present in Latin ( vincere with perfect vīcī '(be) siegen '; '( bes ) siegte ' or 'have (be) won').

The subsequent languages ​​of the Indo-European original language show four different aorist formations, the root aorist, the -s-aorist, the thematic aorist and the reduplicated (also thematic) aorist. Besides the root aorist, the only aorist formation belonging to the original language is the -s-aorist (cf. for example the -s-aorist in Vedic = -σ-aorist in Greek = -s-perfect in Latin). The * -s- goes directly to the root. Without theme vowel, d. H. athematical, the secondary endings follow. In the active, the root is continuously in the-D expansion stage, in the medium, however, in the zero stage, with roots on -y and -w in the -é full stage. Based on findings from Tocharian and Hittite, it is disputed whether the s-suffix is ​​original in all persons or initially only the 3rd Sg. belongs (to the Hittit. * -s- in the 3rd sg of the past tense of the ḫi verbs, but see the contribution here under Anatolian languages ). The presence of an augment is limited to Greek-Armenian-Indo-Iranian and is therefore also questionable for the other individual languages, insofar as they (still) have the aorist.


The perfect stem usually only consists of the reduplicated root. * E is usually used as the vowel of the reduplication syllable (in Vedic also * ē, * i and * u, in Latin also * u, once parallel to each other in ved. Tutóda ~ Latin tutudī , ' abutted ', both probably * stu-stówd - / stu-stud-´), the root is in the active singular in the -ó level, otherwise in the zero level. In Latin, the reduplication perfect has often survived, alongside * stu-stud-´ z. B. still with dare 'give', perfect de dī from the “weak” partial stem * de-dh₃-´, or with cadere 'fall', perfect ce cidī from * k̑e-k̑ód-h₂e + y (~ ved. Glbd. śaśā́da ). A not very frequent exception due to the lack of reduplication is the very old formation 1. Sg. * Wóyd-h₂e 'I know', 1. Pl. * Wid-mé 'we know' from the root * weyd ('see', originally actually 'to find', compare Latin vidēre ' to see') represent (see also preteritopresentia ).

The verb in the subsequent speech

The verb system shown here is most clearly found in Vedic and Greek. This is no wonder, as the reconstruction of Urindo-European is based primarily on these two languages ​​(so-called Graeco-Aryan reconstruction model). The validity of this approach has been questioned; so far, however, no alternative model has been delivered.


The Anatolian languages ​​are believed to have split off prior to the formation of most of the "Graeco-Aryan" features. The - best known - hittit. Verbal system is characterized by the fact that it has given up the aorist and - in contrast to the other individual languages ​​- has not yet developed the perfect. This makes the verbal system much simpler; there is present (expressed by the Present ) and past (= Präteritum expressed by the past tense ), active and passive Medio (the medium has taken over the functions of a passive voice). Thematic verbs play almost no role. Verbs with -o-vowelism ( malli 'mahlt' * mél-molh₂-e + y, dāi 'takes' * déh₃-e + y), the -šša - / - šš- imperfectives (* -sóh₁ - / - sh₁- ´), the -aḫḫ- factitiva (* -eh₂-) and the -anna - / - anni- durativa (* -n̥h₂-óy - / - n̥h₂-i-´; after Kloekhorst-2008, pp. 175f. * - otn-óy- / otn-i-´) are grammaticalized to semideponents with tripod endings in the singular and factive endings in the plural (= ḫi conjugation). All forms of the 3rd pl. Of the past tense receive the tripod ending * r̥, all forms of the 3rd sg. Of the past tense of the ḫi conjugation have the ending * -st, most likely transferred from the very common * h₁és-t 'he was'. The tripod verbs give up their original paradigmatic ablaut -é level: zero level and carry out (like the Vedic and Greek) either the -é level or the zero level in the whole paradigm ( ki-tta (ri) 'he lies' = ved . śáye = Greek κεῖται keítai * k̑éy-e + y or šupp (tt) a (ri) 'he sleeps' * sup- (t) ó ± ri ~ ved. duhé 'she gives milk' * dʰug̑ʰ-é + y; all forms with regular re-formation of the ending). An archaic characteristic is that Hittite can do without the future tense and, with the exception of the imperative, without modes.

In the professional world, it is now (see above) more than certain that the Anatolian language group was the first to leave the general association of speakers of the Indo-European original language with a long time lag. There are too many features that Anatolian does not have but that all other language groups have, e.g. B. the nominal and verbal thematization, the perfect, the modes, the dual, the -tó participle, the -yos comparative or the fact that the -nt participle is a passive participle. Uranatolian is thus actually a sister language of Urindo-European with an unusually large number of linguistically extremely important archaisms, among them the unique phonetic preservation of * h₂ and * h₃ as -ḫ- / -ḫḫ- , and the historically sensational fact that - in the phoneme inventory Luwian and Lycian - all three tectal rows (palatal, velar and labiovelar) can still be distinguished (Kloekhorst-2008, p. 17f.).


The Tocharian language group evidently moved east very early. Tocharisch has diverse, drastic and otherwise non-existent innovations, e.g. B. a system of seven secondary cases, the group inflection, a separate number paral to denote natural pairs (as opposed to the dual, which denotes the numerical duality), a fundamental verbal stem opposition normal verb: causative, and a thematization that starts with the ending * -o (the 3rd part of the tripod endings).


The functions of the various verb forms are most clearly defined in Greek. In addition to the present tense stems (with imperfect tense), aorist and perfect tense (with past perfect tense), there is a future tense stem (with future tense exactly in the passive voice ), which is often, but not always, characterized by an s suffix. The fully developed stock of forms of the perfect medium disproves - together with the Vedic; In both branches of language, which are considered very archaic, the perfect medium is reconstructively identical - the view that a perfect medium came about late, if at all. Decisive for assessing the position of the perfect medium is the groundbreaking and correct assessment by Jasanoff, " the perfect evidently originated within PIE as a kind of ... reduplicated present ". This means that the partial trunk inventory of the present tense was used a second time (with all types of reduplication) and provided with the tripod endings in order to precisely achieve the present resultative meaning of the perfect: Result ( tripod ending) of a completed action ( present tense ). This made the present medium and perfect medium formally identical (since the present medium already had the tripod endings). For the purpose of differentiation, the Greek regulates the distribution of the reduplication vowels as follows: present tense always - i - (Vedic both - i - and - e -), perfect always - e - (Vedic both - i - and - e -) , and aorist always - e - (Vedic both - i - and - e -). Greek 3.Sg. τί-ϑε-ται tí-the-tai means (present tense) 'is set', τέ-ϑε-ται té-the-tai (perfect) 'has been set'. The Vedic does not differentiate here via the reduplication vowel, but via the ending (3rd so-called dhatté 'is set' * dʰe-dʰh₁-téi opposite dadhé 'has been set' * dʰe-dʰh₁-éi) or via the hyphenation of the laryngal ( 2.Sg. dhatsé ‚are set ', dadhiṣé ‚ have been set', originally identical * dʰe-dʰh₁-séi). If both are not possible, the verbal form remains the same: 1.Sg. dadhé 'am set' and 'have been set' * dʰe-dʰh₁-h₂éi.

The active and medium diatheses are joined by a formally differentiated passive in the aorist and future tense. In the present, imperfect, perfect and past perfect, the medium continues to express the meaning of the passive. The renewed differentiation in the aorist and future tense is based on a univerbation with the aorist of the root * dʰeh₁ 'do, do' in the Narten form (Narten = technical term for the addition of a more in both the strong and the weak sub-stem), i.e. strong * dʰḗh₁, weak * dʰéh₁; ἐ-παιδεύ-ϑη-ν e-paideú-thē-n 'was educated' actually means 'was + educated + made'. Since the passive aorist has the active endings, i.e. is a Statofaktiv verb (cf. above in the subsection "Aspect"), he becomes identical (only in the singular) with the active aorist of the verb ( * ἔϑην * é-thēn = ved. Á- dhām * (h₁) é * dʰéh₁-m); for differentiation, the active aorist (only in the singular) is transformed into a - k aorist ( ἔϑηκα é-thē-ka ).

In Greek, athematic conjugation has already lost some ground in favor of thematic.

Vedic and Sanskrit

In Vedic, which has many exact equivalents in Uriran, the variety of forms is even richer than in Greek. However, the nuances of meaning are clearly on the decline. The difference between active and medium is often almost impossible to grasp. However, semantically clear passive forms emerge (a * -yó passive with tripod endings and a passive orist whose origin and ending are not fully understood only in the third section with a -ó-step root and the ending -i ( ákāri 'was made') * (h₁) é * kʷór-i; even without augment jáni 'was born' * g̑ónh₁-i) Even the aspect differences are often no longer recognizable in the Rigveda.

In the niche of an only Indo-Iranian action type category iterative-intensive, an archaic formation is able to survive, which has an athematic present stem from an aorist root by direct doubling of this aorist root ( jáṅ + gan + ti 'comes' * gʷém + gʷom + ti). This formation shows the origin of the root-like * -ó- in hethit. malli 'grinds', thematized in Latin glbd. molō = dt. mahle , with reduplication that is regularly omitted in these individual languages ​​and the following accentuation of the * -ó- from * mél-molh₂ (Vedic mármartu 'should crush').

In later classical Sanskrit the imperfect, perfect and aorist are used as past tenses without any difference in meaning. Verb forms have also been added in Sanskrit: a future tense (also with an s suffix), a passive (here with medial endings and without connection with Greek) and a number of productive derived verb forms such as desiderative or causative. The old subjunctive is only preserved in the forms of the "imperative of the first person".


In the Italian languages ​​(for example Latin) the conjugation system has been heavily rebuilt using the existing building blocks; the result is a more symmetrical and more transparent system.

The athematic verbs have disappeared (with the exception of a few verbs from the basic vocabulary, see above). The thematic verbs were formed by incorporating various suffixes into the known conjugation classes (a, e, “consonantic”, i). For example, verbalization of nouns ending in -a ( cūrāre 'care' from cūra 'care'), a factual * eh₂ suffix ( novāre 'renew' from * new-eh₂-), or an intensive suffix ( canere > cantāre ' to sing'). The ē conjugation is based on a causative suffix * -é-ye- ( monēre ' to warn' from * mon-é-ye- 'to make you think') and a tripod suffix * -éh₁-ye- / * -eh₁-ye- ( alb-ē-re ' to be white', sed-ē-re ' to sit') back. The ī conjugation is based on a series of suffixes and verbalization of nouns on -i- and -o-. Finally, the consonant conjugation continues the thematic conjugation of Urindo-European.

The medium has turned into a passive voice. Of the three aspect systems, the perfect and the aorist have merged to form the perfect system. There are formal elements of the old perfect (endings, occasional reduplication) as well as the aorist (in the -s-perfect, for example dūcō - dūxī , I lead '-' I led 'or' I led '). Both aspect functions can be found, both the perfective aspect ("prematurity", dt. Therefore more like 'I led') and the resultative ("result of a completed action" dt. Therefore more like 'I led').

The tense is now separated from the mode. The old imperfect tense has been lost without a trace. A new past tense with the suffix -bā- takes its place (* bʰwéh₂- 'to be, to be'). A future tense is formed from the old subjunctive with the full step of the root and the theme vowel * -e - / - o- (in the thematic verbs and in the ī conjugation doubled to form * -e- + * -e- = * -ē -). The verbal paradigm is completed by crossing these formations: The future tense erō , I will be, receives a new past tense eram , I war, from the suffix -bā- , the past tense ending in -bā- receives the thematic ending of the future tense -ō , etc. and thus a renewed -bō- future tense for the ā- and ē-conjugation.

The subjunctive goes back (in some of the forms) to the old optative.

Tense, mode, aspect can be combined, but there is no subjunctive in the future tense.


The reconstructive findings of the Germanic strong verbs, which have been undisputedly valid for 200 years, have undergone radical changes and modifications in the direction of a higher degree of correspondence with the verbal relationships of the other individual languages ​​(Mailhammer-2007 in the title: "New System " ). Germanic basic verbs such as bite or pour often find -n-infected equivalents in other individual languages, e.g. B. to bite lat. Findō “I split” and to pour lat. Glbd. fundō . The assumption that * bʰid-ó- or * g̑ʰud-ó- is the common starting point for on the one hand (German.) * Bʰ + e + yd-ó- / * g̑ʰ + e + wd-ó- and on the other hand (Latin) * bʰi + n + d-ó- / * g̑ʰu + n + d-ó- is supported by the - extremely rare - presence of root-like verbal zero levels in Gothic digan 'knead' * dʰig̑ʰ-ó- (Latin fingō , I form ' * dʰi + n + g̑ʰ-ó-) and ais. vega 'fight' * wik-ó- (Latin vincō , I win '* wi + n + k-ó-). Important philological preparatory work in Seebold-1970 also shows that the Germanic strong verb has the vowelism of the "type bhárati " (i.e. the stressed -é full stage of the root), but the consonantism of the "type tudáti " (that is, the zero stage of the root with stressed theme vowel) ) (Mailhammer-2007, p. 133: ... significant discovery ... with references to the effects on traditional doctrine). The root vowelism scheme of strong verbs would not be - using the example of the second strong verb class - (Pres.) * Éw (Prät.Sg.) * ów (Prät.Pl.) * u (Pz.Prat.) * U, but ( in the order given) * u - * ów - * u - * u. The Germanic strong verb in the present stem would not be basic -é-stage and "proterokinetic", but zero-stage and "hysterokinetic".

Kroonen-2013 adds an athematic * -néh₂ - / - nh₂-´-Intensivum (with a zero step root) to the traditional series as practically regular. Together with the causative-iterative formation on * -é-yo-, each strong verb would then have six forms of realization, i.e. to the root * dewk 'draw': * déwk-o- * de-dówk- * de-duk-´ * duk -O- * dowk-E * duk-néh₂ - / - NH₂-'(. dt prefer * ZOCH subjected ge-coated witness twitch / pull out ), or the root wreyd *, scores': * * we wrid-O- * we -wróyd--wrid-'* wrid-O- * wroyd-E * wrid-néh₂ - / - NH₂-' (. dt tear * tear cracks ge-torn irritate scratch ). Not everyone always educates German. Individual dialects completely make up the series of forms, but are German throughout the whole. Dialect area such complementary examples very numerous.

Compared with the Hittite verb, which is considered to be very original, the Germanic additionally only has the perfect (the only non-periphrastic past) and the * -yéh₁ - / - ih₁-optative (which develops into the subjunctive). In Anatolian, other categories such as the thematic subjunctive or the Graeco-Aryan variety of forms are considered “not yet established”, but in Germanic the verbal system is considered to be “greatly simplified”. It may be revolutionary, but it is obvious that Germanic behaves more like Anatolian in this respect.

Verbs for which no inherited perfect originally existed are called weak verbs. They form their past with a new suffix * -d-, which most likely goes back to the perfect tense of the verb tun (* dʰe-dʰóh₁- / dʰe-dʰh₁-´).

A mediopassive is still preserved in Gothic, but ultimately in this form - with the exception of a few remains, e.g. B. in Old English - extinct in all Germanic languages. Passive forms are newly formed periphrastically, and many other forms are replaced by periphrastic formations (auxiliary verb constructions) , as in many other subsequent languages .


In the Slavic languages, aspect is expressed lexically . The concept of the aspect (as a view of the state of affairs, in contrast to the type of action as a kind of state of affairs) originates from the study of the Slavic languages.

Sentence syntax

Less clear statements can be made about the sentence structure of the original language than about the theory of forms, since a means such as the analysis of the phonetic / phonological developments that typically behave very regularly, from which one can draw conclusions about the morphology, is not available at the sentence level Has. It remains to collect typical sentence patterns of the early forms of the subsequent languages ​​and carefully draw conclusions as to the extent to which these could already have existed in the Indo-European original language.

We are used to German that a main clause contains at least one subject and one predicate. In Latin, for example, it is different: Here a first or second person pronoun may only be used if it is stressed, so that sentences arise without a formal subject. This situation is also assumed for the original language. However, through the verb form we still have an implicit subject predetermined by person and number; Incidentally, in some non-Indo-European languages ​​even that is not required.

Complete sentences with a purely nominal predicate were also common: the copula , which connects the subject and the predicate noun as a formal verb (the man is beautiful; the woman is a craftsman; mother is at home), for example, does not occur in modern Russian . It is assumed that such nominal sentences (man beautiful, woman craftswoman, mother at home) were common in Indo-European. The verbs * h₁es- (exist), * bʰew- (become) and others appear in the following languages ​​as (often optional) copula (cf. he is , I am ).

The verb was usually at the end of the sentence, but any part of the sentence could be dragged to the beginning of the sentence to emphasize it (Latin habent sua fāta libellī , es have their fates have the books, German requires the “es” in front of the verb). In the island Celtic languages, verb confrontation has become the standard.

Syntactic relationships between nouns, adjectives, pronouns and verbs were established by congruence of the inflected forms.

For the organization of sentences and sentence sequences serve enclitics : trailing particles (or inflected words) whose accent then transferred to the front standing word. Examples are the Latin -que (= Greek -τε , Vedic -ca , Indo-European * -kʷe), Greek sentence structure particles such as μέν… δέ - mén… dé 'although… but', or the enclitic pronouns.

Such enclitics are particularly popular in the second position of the (main or partial) sentence ( Wackernagel's law ). Chains of enclitic particles at this point are particularly typical of the Hittite.

Question sentences are identified by the use of question pronouns or question enclitics (for example, Latin -ne ), negation by the adverb * ne and the word prefix * n̥-.

Relative clauses use the relative pronoun and precede the main clause. It is assumed that in the original language, as in Sanskrit, these did not refer directly to the nouns, but to separate demonstrative pronouns in the main clause. (In German this difference is somewhat blurred by the articles; in Latin the opposite situation exists that relative clauses, both as subject and attribute clauses, do not need a reference pronoun.) The two types of relative pronouns (* kʷí - / * kʷó- and * (h₁ ) yó-) correspond to the two types of relative clauses (explicative and restrictive).

Other types of subordinate clauses, for example causal clauses introduced by conjunctions, cannot be reconstructed.

In the following conversations one knows an absolute participle construction, for example the Latin ablative absolutus, the Greek genitive absolutus, the old Indian locativus absolutus or the old church Slavonic dative absolutus. It is not entirely clear whether these constructions are based on a common grammatical structure or whether they are innovations in individual languages. The original construction was most likely (also semantically obvious) the one with Locativus absolutus (so taken up again in modern languages, e.g. English with things being the way they are , German "with the traffic light switched off"). The linguistic distribution of the construction is most plausibly due to the respective case syncretism.


In the area of ​​the basic lexicon, the very extensive collection of material by Pokorny (1959) is unsurpassed to this day (examples in the article Indo-European word roots ). Except phonetically inevitable schwas however Pokorny does not provide laryngeals; however, these are usually easy to add.

Word formation


The Urindo-European, like all languages, probably adopted words from other languages. Today, however, there are no known examples of words that were clearly borrowed from namable neighboring languages ​​in Urindo-European times. However, some words are very likely loanwords due to their atypical shape; Well-known examples are * h₂éb-ōl 'apple' or * angh₁-lo- (for example) 'messenger of the gods' (from Greek ἄγγελος in German 'angel'). Another is * peleḱus 'ax' (cf. ancient Greek pélekys , osset. Færæt , Skt. Paraśú ), which was previously associated with the Akkadian pilakku until its meaning as 'spindle', not 'ax', emerged .


The most important means of word formation from roots and other words were the suffixes already mentioned.

The table shows some important word formation suffixes :

suffix meaning Examples
* -yo- Affiliation (adjectives) lat. pater 'father' - patr iu s 'fatherly'
* -ey-o- Substance (adjectives) Latin aurum 'gold' - aur eu s 'golden'
* -tó-, * -nó- Past participle passive,
verbal adjectives
* ǵerh₂- 'grind' → * ǵr̥h₂- - ' ground up ' → lat. grā nu m , aslaw. zrŭ no , got. kaúrn , all 'grain';
in German also z. B. * stoih₂-nó-s 'solidified' → German stone , or * dr̥-nó-s 'torn' → German anger
* -ih₂- Feminine education * dei-w-ó-s → ved. devás 'God', * dei-w-íh₂ → ved. dev ī́ 'goddess'
* -eh₂- * dei-w-ó-s → lat. deus 'God', * dei-w-éh₂ → lat. de a 'goddess'
* -ḱo- reduction * h₂i̯u-h₁n̥-ó- 'young' (cf. ved. yúvā , lit. jáunas ) → * h₂i̯u-h₁n̥- ḱó s → ved. yuvaśáḥ 'youthful', Latin iuvencus 'young bull', German young
* -lo- reduction lat. - (u) lu-s, e.g. B. in * dwé-no-s (→ Latin bonus 'good') → * dwé-ne -lo- s → Latin bellus 'beautiful'
* -teh₂- Abstraction * néwo-teh₂-ts → Latin novi s 'novelty'
*-goal- Perpetrator Latin orā tor 'Redner' (German -ter in ter is, however, the loan suffix -ārius from Latin)
* -h₂ter- relationship * Méh₂tēr , * ph₂tḗr , * bʰréh₂tēr , * dʰugh₂tḗr → Mu tter , Va ter , Bru of , Toch ter
- but not Sister ter because of * swésōr arose or El ter n (from the comparative elders )
* -tro- Instrumental formation * h₂erh₃- 'plow' → * h₂erh₃- tro - 'means of plowing'
→ Greek ἄρο τρο ν áro tro n , Latin arā tru m , poor. arawr , welsh.aradr , me. arathar , order. arðr , all 'plow'

Accent / ablaut shift

The change from one accent / ablaut class to another was a word formation tool. An example from a proterokinetic inflection is * bʰléǵʰ-men- (holy word, cf. skr. Bráhmaṇ - ), from an amphikinetic or hysterokinetic inflection * bʰleǵʰ-mén- 'priest' (skr. Brahmán- with Vr̥ddhi brāhmaṇa - ).

A variant of the word-forming use of the ablaut in nouns that only occurs occasionally has become very productive in Sanskrit: the so-called Vṛddhi formation. Here, a derived noun is formed from a basic noun by bringing the root syllable into the expansion stage. Examples are known from religious terminology: A follower of the god V i shnu is a V ai shnava (in Sanskrit ai is the extension of i ), a follower of the Sh i va a Sh ai va, a follower of the J i na a J ai na (hence the two terms Jinism and Jainism for this religion).


Word formation through composition , as it is also typical for New High German, is also assumed for Urindo-European, but to a much lesser extent than later in Greek or even in Sanskrit. Nouns were attached to one another, the hind limb was inflected. Noun compounds were not common in all subsequent languages, they are rarely found in Latin, and practically none at all in Hittite.

Linking with adverbs and prepositions led to the verbal prefixes of the subsequent speeches.

Typical are personal names (* h₁néh₃-mn̥-), which are made up of two religiously / socially significant components: Greek Themisto-kl (s (law fame), Old High German Ans-elm (God helmet), Czech Bohu-slav (God fame ), Gallic Catu-rīx (battle king) or Irish Fer-gus (hero power).

Vocabulary analysis

From the common vocabulary one tries to draw conclusions about the civilization and culture of the language community of Urindo-European. An important example is the tribe * kʷ-kʷlh₁-ó -, * kʷe-kʷlh₁-ó-, * kʷé-kʷlh₁-o- or * kʷó-kʷlh₁-o-, which us in the meaning of 'wheel' (or in similar , meanings derived from it) is handed down in many subsequent languages:

urindo-european * kʷ-kʷlh₁-ó-:

→ Hittite kugullaš

urindo-european * kʷe-kʷlh₁-ó-:

→ Vedic cakrá-
→ avestisch čaxra-
→ tocharisch A kukäl 'carriage'
→ Tocharian B kokale 'wagon'
→ Old English hweowol, hweogol

urindo-european * kʷé-kʷlh₁-o-:

→ ancient Germanic * hweh (w) ulaz
→ Old Norse hvēl, hjōl
→ Icelandic hjól
→ Old English hwēol
→ English wheel

urindo-european * kʷó-kʷlh₁-o-:

→ Greek κύκλος kýklos 'circle'
→ Slavic kolo
→ Lithuanian kãklas 'neck'

The etymology of this word stem can also be explained: It is obviously a reduplication of the verbal root * kʷelh₁ (meaning 'to turn'), which here in its themed "weak" sub-stem * kʷlh₁-kʷlh₁-ó- with the reduplicative variants * kʷ - / * kʷe- / kʷé- / kʷó- is realized, whereby the order probably corresponds to a chronological sequence. This doubling is semantically an iconic representation of the repeated turning movement of the wheel.

Since a common word for both 'house' and 'door' can be reconstructed from the subsequent conversations, one can assume that the speakers of the Indo-European original language were already settled, i.e. not nomads.

Individual language examples of the word for house are:

urindo-European * dem (with a highly archaic gen. * dém-s) in the individual language-typical ablaut levels and with an -u extension of the stem * dom-ú-:

→ Vedic Gen.Sg. dán * dém-s = Greek δεσ- des- in δεσπότης des-pótēs 'master of the house'
→ Vedic dámaḥ * dém-os
→ Greek δομός domós * dom-ó-s
→ Armenian town [tūn] * dṓm-s, Gen. tan * dm̥-és
→ Latin domus * dom-ú-s = Old Church Slavonic domŭ

Individual language examples of the word for 'door' are:

Urindo-European * dʰwer (originally only in the dual, reference to "door wing"; often in the plural in individual languages):

→ Vedic Nom.Pl. dvấraḥ * dʰwḗr-es (* -ṓ-? * -ó-?), with loss of aspiration probably after dvấ 'two'
→ Greek θύρᾱ thýrā * dʰúr-eh₂
→ Armenian Nom.Pl. durkʿ * dʰúr-es
→ Latin Nom.Pl. forēs * dʰwór-es
→ Gothic daúr , Old High German tor , New High German Tor * dʰur-ó-m (further training for the -i tribe in German Tür )

Overall, the reconstructed vocabulary suggests a Neolithic agricultural culture that knew milking, cows, sheep and horses. A particularly important argument for this hypothesis is the reconstruction of the verb 'plow' (which, however, is missing in the Indo-Iranian languages):

simply * h₂erh₃- (cf. Tocharian AB āre ‚plow '* h₂érh₃-os, * h₂érh₃-om or (neutral -s-stem) * h₂érh₃-os):

→ Hittite ḫarrai ‚tears open , crushed '* (h₂ér-) h₂orh₃ (archaic intensive, always → ḫi conjugation)
→ Greek ἀρόω aróō (probably zero step * h₂r̥h₃-ó-)

with -yo extension * h₂erh₃-yé / ó-:

→ Lithuanian ariù
→ Old Church Slavonic orjǫ
→ Latin arō (-āre) (probably after arātrum 'plow')
→ Central Irish airim
→ Gothic arjan , old English erian , old high German erien , erran

Furthermore, the verb 'milken' can be reconstructed for the original language, but the reconstruction of 'milk' is controversial.

However , the hints from the vocabulary are not sufficient to clearly determine the original home . A case study of such an analysis is described in the article Salmon Argument .

The kinship terms in the common vocabulary were studied most intensively. Characteristic features here are, for example, that no distinction is made between older and younger siblings, and the strange fact that “nephew” and “grandchildren” are referred to with the same word.

More about the statements about the speakers obtained from the vocabulary can be found in the articles Indo-Europeans , Urheimat , Kurgankultur .

Reconstruction methods

The comparative method

This historical-comparative method (also called comparative method ) was developed in the 19th century on the basis of the Indo-European languages ​​and became the standard method of historical linguistics for the reconstruction of the preforms in language groups. It works best (but not exclusively) in the field of phonology, as sound changes are typically very systematic.

Correspondence rules are formed from potential word equations , which are checked on other relatives and adapted if necessary. On the basis of this regular correspondence, one models plausible original linguistic output forms and phonological plausible development paths from the original forms to the individual linguistic sounds. In this way, you can reconstruct original language roots and grammatical forms.

The possibilities and limitations of this method can be the comparison of the of the Romance languages developed proto-language seen with the traditional Latin. The existence of the Latin h cannot be deduced from the Romance languages, since the sound was already lost in Latin before it was separated into the subsequent languages. Likewise, the synthetic passive of Latin ( laudor, laudāris, laudātur , etc.) is not preserved in any Romance language .

Internal reconstruction

This method only looks at a single language, typically the already reconstructed original language itself. A regularity is found in the language, to which there are exceptions. Based on the assumption that the exception forms were also regular in an earlier language form, the earlier rule system and the change processes that led to the exceptions are modeled.

Relationship calculations: lexicostatistics and glottochronology

Since the beginning of the 20th century, attempts have been made to calculate the relationships between the Indo-European languages ​​and the relationships between them using lexicostatistical methods . The glottochronology also attempts time accepting more or less continuous replacements in a universal list of terms ( Swadesh list ) to close to the age of the different language levels. The relatively constant changes documented in some examples are, however , falsified in many other cases by known or unknown, precisely non- regular, since socio-historical events. Despite multiple attempts to take these falsifications into account, no really convincing results have been presented so far.

Typological procedures

From the observation of many languages ​​in the world, it has been found that certain syntactic properties of languages ​​typically appear together. So Winfred P. Lehmann concluded , based on the word order typology of Theo Vennemann , that in the original language the verb was at the end of a sentence ( subject-object-verb ). On this basis he was able to postulate further syntactic properties of the original language. The approach is controversial: some reject it entirely, others are cautiously benevolent.

Timeline of the history of research

year Researcher Contribution
1814 Rasmus Christian Rask Carry out detailed comparative studies of different Indo-European languages
1816 Franz Bopp
1819 Jacob Grimm
1833-1836 August Friedrich Pott Establishes the Indo-European etymology
1861 August Schleicher Carries out first precise reconstructions and establishes the family tree theory ; its reconstructed original language shows great similarities with Sanskrit
1876 Hermann Osthoff Develops syllabic alveolar cells
1876 Karl Brugmann Opens syllabic nasals
Accepts both voiced and unvoiced aspirates and more fricatives (voiced s as well as ð and þ, and their aspirated versions)
Develops the basics of morphology
1877 Karl Verner Formulates Verner's law, d. H. the influence of the word accent on the phonetic development of certain Germanic plosives
1878 Ferdinand de Saussure No longer adopts the a-vowelism of Sanskrit, but rather eoa as the basic vowels of the original language
Laryngale: He suggests two indefinite vowel-like sounds, which he calls “coefficients sonantiques”
    The doctrine subsequently assumes a single Schwa sound for the coefficients
1880 Hermann Möller Proposes a third coefficient and assumes a laryngeal loudness for all three
1890 Peter von Bradke Adopts a basic dialect division into Kentum and Satem languages. The assumption is accepted doctrine until the late twentieth century (until the discovery of Tochar)
1893,1897 1900 Berthold Delbrück Creates a comparative syntax of the Indo-European languages
1895, 1900 Hermann Hirt Clarifies important questions about the accent and the ablaze
1912 Albert Cuny Describes the essential elements of today's laryngeal theory in an essay
Per Persson Performs the systematic study of the suffixes
1926, 1928 Jacob Wackernagel Examines sentence syntax, especially the role of enclitics
1927 Jerzy Kuryłowicz Identifies the Hittite with the second laryngeal of Cuny
1927-1932 Alois Walde , Julius Pokorny Carry out significant research in the field of the lexicon
1973 Tamas Gamqrelidze , Vyacheslav Vsevolodowitsch Ivanov Beat the Glottalhypothese ago
1974 Winfred P. Lehmann Applies language typology methods to syntax research
up to 197x   The laryngeal theory is not accepted by all researchers until the 1970s
1975, 1976, 1998 Helmut Rix , Karl Hoffmann Shed light on the verb tense, mode and aspect systems

Sample texts

August Schleicher

From the Indo-European original language - a reconstruction - no texts have survived, but attempts have been made to write texts in Ur-Indo-European. "Avis akvāsas ka", an Indo-European fable constructed by August Schleicher in 1868, is particularly prominent to this day (cf. there more recent attempts at transmission); According to the state of the art of linguistics, this fable has been repeatedly adapted to the current hypotheses for the structure of the Indo-European original language. For example, Schleicher's “Avis akvasas ka” became “Owis eḱwōskʷe” in a more recent version from 1979; Perhaps in 2013 one would say “* h₃éw-is h₁ék̑-wo-es-kʷe”, but it will be different again in future years. In linguistics, however, the reconstruction of entire texts is fundamentally very speculative.

For the computer game Far Cry Primal , various constructed versions of the Indo-European original language were set to music.


  1. "wheel" * kʷe-kʷlh₁-ó - / * rót-h₂-o-, "axis" * h₂ég̑-so-
  2. In the "Cappadocian tablets" (19th / 18th century BC) from Kültepe, the old Kārum Kaneš (German for 'Kanisch trading place'), an Assyrian trading colony in Anatolia, written in the old Assyrian dialect of Akkadian a number of ancient ethite forms of name on (male) -ḫšu , (female) -ḫšu-šar , which here mean 'son' or 'daughter'. -ḫšu is the syncopated form of the later heth. ḫaššu- 'king' (literally "child of the whole country"). These name components are not only by far the oldest documented word forms in the language family, but can also be etymologized: ḫaššu- * h₂éms-u- is in the words for 'God' ved. ásura- , avest. ahu, ahura and old iceland . áss (German Ase ); * this 'woman' serves e.g. B. on archaic feminizations of numerals three and four (for further details see the articles Kültepe and Cappadocian Tables ). In the zero level * -sr- there is * this 'woman' in German sister * swé- sr - and cousin from Latin cōnsobrīna * kon + * swé- sr -ih₂-neh₂.
  3. ↑ In terms of developmental history, * bʰor-ó-s or * dʰrogʰ-ó-s is a regularly occurring thematization from the Gen. Ab. * bʰor-ós / * dʰrogʰ-ós of a weak sub-stem, while * bʰór-os and * dʰrógʰ-os are about the secondary theming of a strong sub-stem * bʰór / * dʰrógʰ (i.e. * bʰór + -o- / * dʰrógʰ + - O-). In the case of the root * carry- 'carry' the nom.Sg. the original athematic paradigm * bʰṓr-s in Greek phṓr , thief '(= glbd. Latin for ), which is assumed for the above-mentioned derivations, is still alive.
  4. The possibility of being able to form both a counting plural and a collective plural is a highly archaic linguistic phenomenon, cf. hittit. alpaš 'cloud' * albʰ-os, alpēš 'clouds' * albʰ-ey-es <* albʰ-o-es, and alpa ḪI.A 'cloud' * albʰ-e-h₂ (superscript ḪI.A is the Sumerian symbol for the plural of the "factual class"). The possibility of counting plural images alongside simultaneous collective plural is, however, on the decline in the younger language levels.
  5. cf. next to τὰ ζῷα τρέχει tà zõia tréchei tá zōa tréchei "the animal runs" z. B. still πάντα ῥεῖ pánta rheĩ , German 'everything flows' (attributed to Heraclitus).
  6. The declension classes vary in the individual languages ​​because the basic language is able to add the ablaut vowel * -é- to either the root, the suffix or the ending in the “weak” sub-stem. The difference in the individual language declension classes is due to the fact that the individual languages ​​make their selection in different ways (the unselected forms are then given up). From the options * h₂éw-is, * h₂w-éy-s and * h₂w-y-és z. B. Latin. * h₂éw-is and the Altind. * h₂w-éy-s (old Indian gen . véḥ ' des Vogel') from. From the options * péh₂-wn̥-s, * ph₂-wén-s and * ph₂-un-és z. B. the Hittite. * ph₂-wén-s and the Greek. * ph₂-un-és (Greek Gen.Sg. πυρόϛ ‚des Feuers' instead of * φυνόϛ with the consonantism of the expected Nom.Sg. * πᾶαρ , which conversely receives the quantitatively graduated (i.e. elongated) vowelism of the Gen.Sg., thus appears as πῦρ ). When making the selection, certain regularities can be seen in the individual languages. "Strong" sub-stems are often transformed according to the phonetic characteristics of the "weak" sub-stems (e.g. old Indian nom . Uṣā́s 'dawn' instead of * óṣās ). - Because of its limited phonetic development possibilities and the homophonicity with the "strong" sub-stem, the -é-deformation of the root in the "weak" sub-stem is only rarely found and mostly in archaic forms (* dém-s 'of the house' in Greek. δεσπότης des-pótēs 'master of the house', * négʷʰ-ts 'of the night' in hethit.nekuz mēḫur 'time of the night' or - from the -o root * gʷow 'ox' - * gʷów-s 'of the ox' in Vedic gós ). The preservation of this type of declination class in the Latin. -i trunks is a noticeable specialty.
  7. In Luwischen this * w is also in the 1st Sg. transfer; the singular ending sentence of the mi conjugation is thus hethit. 1.Sg. -mi , 2.Sg. -ši , 3rd sg. -zi , but luw. 1.Sg. -wi , 2.Sg. -si , 3.Sg. -ti .
  8. In Greek Aspirates in endings do not cause a slight dissimilation (e.g. 2nd pl. Aor. Med. ἔϑεσϑε éthesthe 'you put for yourself' * (h₁) é * dʰh₁- (s) dʰwé). The -ϑη- element in the aor. and Fut. Passive, however, triggers breath dissimilation, so that it can also be assumed that it is not an ending but a verbal contraction.
  9. The (uncertainly proven) past tense of the Germanic -yo-verb with -o-step root is formed without the -yo-suffix. This would be one of the cases in which the -ō-vowelism of the Germanic VI. Verbal class (e.g. German drove , created or English shook , shook, took , took) from * h₂é-h₂orh₃> urgerman. * ōr arose according to the law. In the case of preservation, the verb would be called today * ären with the simple past * ur (like swear swore , actually * swear * swore from * swor-yó-, or lift raised , actually * häppen * huf from * kap-yó- = Latin capiō ‚ seize '). The old high is unclear. Past tense 3rd place ierun 'they plowed', that with Urerman. * -ē²- is formed, so clearly indicates the earlier existence of a Geminate in the present stem (urgerman. * arr-ja-). The Geminate might go to the Hittite. well-known archaic image Sg. * h₂ér-h₂orh₃, Pl. * h₂r-h₂rh₃-´ back. The plural form is able to generate a transitional sound (ie * h₂r-h₂rh₃- [r] -´) before the vowel-like ending, which has become fixed and transferred to the entire paradigm and thus to a present stem urgerman. * arr-ja- led to the formation of the simple past with * -ē²-, but was then given up again.
  10. * né-pōt-s 'not the Lord' = 'like the Lord', perhaps indicating the similar appearance of close relatives, or semantically parallel to hierogl.-luw. nimuwiza 'child' to muwa 'strength' (etymology uncertain), actually that which has no strength yet ?
  11. Improving or clarifying developments in the notation area could be: Notation of a “strong” sub-stem always without a hyphen, a “weak” one always with ; Designation of a suffix or infix with a double hyphen; Specification of a zero level through an expressly designated space. The notation would then be "* h₃éw = i -s * h₁ék̑ = wo- es + * kʷe". Lat. The Nom Sg: NB.. Ovis , sheep 'would have as reconstructions of * h₃éw = i -s, Gen. Sg. Latin ovis 'of the sheep' but * h₃éw = i- s. The notation method gives clear indications of nominal (and verbal) stem formation characteristics from the reconstructive statement and information using the simplest means.


  • David W. Anthony, Don Rings: The Indo-European Homeland from Linguistic and Archaeological Perspectives. In: Annual Review of Linguistics. Issue 1 (2015), pp. 199-219 (online) . (English)
  • Robert SP Beekes: Vergelijkende taalwetenschap. An inleiding in de vergelijkende Indo-europese taalwetenschap . Het Spectrum, Amsterdam 1990. (Dutch)
    • English: Comparative Indo-European Linguistics. An Introduction. 2nd Edition. Translated by UvA Vertalers / Paul Gabriner. John Benjamin, Amsterdam / Philadelphia 2011 (1st edition 1995).
  • Wilhelm Braune, arr. v. Hans Eggers: Old High German Grammar. 13th edition. Niemeyer, Tübingen 1975, DNB 750227532 .
  • Andrew Byrd: The Indo-European Syllable . Brill, Leiden 2015. (English)
  • James Clackson: Indo-European Linguistics. An Introduction . Cambridge University Press , Cambridge 2007, ISBN 978-0-521-65313-8 . (English)
  • Will Chang, Chundra Cathcart, David Hall, Andrew Garrett: Ancestry-constrained phylogenetic analysis supports the Indo-European steppe hypothesis. In: Language. Volume 91, Issue 1 (2015), pp. 194–244 (online) . (English)
  • Benjamin W. Fortson IV: Indo-European Language and Culture. An Introduction. 2nd Edition. Blackwell Publishing, Malden 2010, ISBN 978-1-4051-0316-9 . (English)
  • Matthias Fritz: The dual in Indo-European . University Press Winter, Heidelberg 2011.
  • Tamaz V. Gamkrelidze , Vjačeslav V. Ivanov : Indoevropjskij jazyk i indoevropejcy. Reconstruction i istoriko-tipologieskij analiz prajazyka i protokultury. University Press Tiflis, Tiflis 1984. (Russian)
    • English: Indo-European and the Indo-Europeans. A Reconstruction and Historical Analysis of a Proto-Language and a Proto-Culture . 2 volumes. Mouton de Gruyter, Berlin / New York 1994–1995.
  • Karl Hoffmann: Essays on Indo-European Studies. Volume 1, Wiesbaden 1975, ISBN 3-920153-47-2 , Volume 2, Wiesbaden 1976, ISBN 3-920153-51-0 .
  • Jay H. Jasanoff: Hittite and the Indo-European Verb. Oxford University Press, 2003, ISBN 0-19-928198-X . (English)
  • Ernst Kausen : The Indo-European languages. From prehistory to the present. Helmut Buske Verlag, Hamburg 2012, ISBN 978-3-87548-612-4 .
  • Alwin Kloekhorst: Etymological Dictionary of the Hittite Inherited Lexicon. Brill, Leiden / Boston, 2008, ISBN 978-90-04-16092-7 . (English)
  • Guus Kroonen: Etymological Dictionary of Proto-Germanic. Brill, Leiden / Boston 2013, ISBN 978-90-04-18340-7 . (English)
  • Winfred P. Lehmann : Theoretical Bases of Indo-European Linguistics. Routledge, London / New York 1993, ISBN 0-415-13850-7 . (English)
  • Arthur A. Macdonell: A Vedic Grammar for Students. 11th edition. Oxford University Press, 1987, ISBN 0-19-560231-5 (English)
  • Robert Mailhammer: The Germanic Strong Verbs, Foundations and Development of a New System. (Trends in Linguistics, Studies and Monographs 183). Mouton de Gruyter, Berlin / New York 2007, ISBN 978-3-11-019957-4 .
  • JP Mallory, DQ Adams: The Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World . Oxford University Press, Oxford / New York 2006. (English)
  • Michael Meier-Brügger: Indo-European Linguistics. 9th edition. de Gruyter, Berlin 2010, ISBN 978-3-11-025143-2 .
  • Holger Pedersen: "On the question of the original relationship between Indo-European and Ugro-Finnish". In: Mémoires de la Société finno-ugrienne. 67, 1933, OCLC 177284500 , pp. 308-325.
  • Holger Pedersen: Hittite and the other Indo-European languages. (Danske Videnskabernes Selskab, historisk-filologiske Meddelelser, 25/2). Munksgaard, Copenhagen 1938.
  • Asya Pereltsvaig, Martin W. Lewis: The Indo-European Controversy. Facts and Fallacies in Historical Linguistics . Cambridge University Press, Cambridge 2015. (English)
  • Julius Pokorny: Indo-European etymological dictionary. Francke Verlag, Bern / Munich, Volume I, 1959, Volume II 1969, DNB 457827068 .
  • Eduard Prokosch: A comparative Germanic Grammar . Linguistic society of America, University of Pennsylvania, Philadelphia 1939. (New edition: Tiger Xenophon 2009, ISBN 978-1-904799-42-9 )
  • Don Rings: From Proto-Indo-European to Proto-Germanic. Oxford University Press, Oxford / New York 2006, ISBN 0-19-928413-X . (English)
  • Helmut Rix : Historical grammar of Greek. Phonology and form theory. Darmstadt 1976, ISBN 3-534-03840-1 .
  • Helmut Rix, arr. v. M. Kümmel, Th. Zehnder, R. Lipp u. B. Schirmer: Lexicon of Indo-European Verbs . The roots and their primary stem formations (LIV). 2nd Edition. Wiesbaden 2001, ISBN 3-89500-219-4 .
  • Elmar Seebold: Comparative and Etymological Dictionary of Germanic Strong Verbs. Mouton, The Hague 1970, DNB 458930229 .
  • Andrew L. Sihler: New Comparative Grammar of Greek and Latin . Oxford University Press, Oxford / New York 1995, ISBN 0-19-508345-8 .
  • Oswald JL Szemerényi: Introduction to Comparative Linguistics. 4th revised edition. Scientific Book Society, Darmstadt 1990, ISBN 3-534-04216-6 .
  • Eva Tichy: Indo-European basic knowledge . Hempen Verlag, Bremen 2000, ISBN 3-934106-14-5 .
  • Karl Verner: An exception to the first sound shift. In: Journal for Comparative Linguistic Research. 23, 1877, pp. 97-130.
  • Mariona Vernet i Pons: La Segona Conjugació verbal llatina: Estudi etimològic i comparatiu sobre l'origen protoindoeuropeu de la formació dels seus temes verbals . Barcelona 2008, ISBN 978-84-477-1030-0 (Catalan)
  • Harald Wiese: A journey through time to the origins of our language. How Indo-European Studies explains our words. Logos Verlag, Berlin 2007, ISBN 978-3-8325-1601-7 .
  • Dagmar S. Wodtko, Britta S. Irslinger, Carolin Schneider: Nomina in the Indo-European Lexicon. (Indo-European Library. Series 2: Dictionaries). Heidelberg 2008, ISBN 978-3-8253-5359-9 .
  • Ingo Zahn: Comparative Indo-European theory of forms. Publishing house Dr. Kovač, Hamburg 2014, ISBN 978-3-8300-7648-3 .

Web links

Individual evidence

  1. ^ JP Mallory, DQ Adams: The Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World. Oxford University Press, Oxford 2006, ISBN 0-19-929668-5 .
  2. Fortson, 2.58f
  3. Fortson, 2.73f
  4. after Fortson, 7.14
  5. The second youngest period of the Holocene in North-West Europe is Subboreal (3710-450 v. Chr.), It is also called Late heat time or mixed oak forest referred -Erlenzeit.
  6. Meier-Brügger E507
  7. Winfred P. Lehmann: Proto-Indo-European Syntax. Univ. of Texas Press., Austin 1974, ISBN 0-292-76419-7 .
  8. Meier-Brügger, chap. II; Fortson chap. III
  9. Lehmann 1966, 5.2.2 last paragraph
  10. Kroonen 2013, p. 383.
  11. cf. Euler 2009, p. 79.
  12. ^ Donald Rings: From Proto-Indo-European to Proto-Germanic. A Linguistic History of English. v. 1. Oxford University Press, Oxford 2006, ISBN 0-19-928413-X , p. 60.
  13. Meier-Brügger F214
  14. Damaris Nübling u. a .: Historical linguistics of German. (= Fool study books). Tübingen 2006, ISBN 3-8233-6212-7 , 9.1.2 (Table 24).
  15. ^ Studies on the morphophonemics of the Indo-European basic language. P. 136 ff.
  16. Meier Brügger F 304 (7th edition)
  17. ^ Wilhelm Braune, Frank Heidermanns (arr.): Gothic grammar. 20th edition. Max Niemeyer, Tübingen 2004, p. 142.
  18. ^ Forms from:
    EV Gordon, AR Taylor: An Introduction to Old Norse . 2nd Edition. Clarendon Press, Oxford, p. 293.
  19. ^ After Karl Brunner: Old English grammar. 3. Edition. Max Niemeyer, Tübingen 1965, p. 259.
  20. a b The Gen. You. the 1st person (* ugkara ) and the nom. Du. the 2nd person (* jut ) are not documented, but can be reconstructed.
    see: Braune / Heidermanns, 2004, p. 132f.
  21. see: Braune / Heidermanns, 2004, p. 132f.
  22. The "g" in the letter combinations "gk" and "gq" denotes a velar nasal [⁠ ŋ ⁠]
  23. ^ Wilhelm Braune, Ingo Reiffenstein: Old High German Grammar I. Phonology and Forms. 15th edition. Max Niemeyer, Tübingen 2004, p. 241.
  24. Braune / Reiffenstein, 2004, p. 182.
  25. The old Icelandic plural pronouns are used for polite salutation or in high style.
    see. about this: Stefán Einarsson: Icelandic . Johns Hopkins University Press , Baltimore / London 1945, reprint: 1994, p. 68 u. 122
  26. Fortson chap. 17th
  27. Meier-Brügger F303
  28. Topic 12 Historical Morphology: The Noun. Grammatical categories. Word formation. Paradigm. ( Memento from October 19, 2014 in the Internet Archive )
  29. Fortson 2004, 2nd edition. 2010, p. 103: " The perfect ... endings ... closely resemble those of the middle ".
  30. ^ Meier-Brügger, F200
  31. Jasanoff 2003 in detail pp. 7–17.
  32. Jasanoff 2003, p. 169.
  33. Fortson, October 5
  34. Guus Kroonen: Etymological Dictionary of Proto-Germanic . Brill, Leiden / Boston 2013, p. 31f.
  35. ^ Hans J. Holm: Genealogical relationship. In: Quantitative Linguistics. An international manual . (= Handbooks for Linguistics and Communication Studies. Volume 27). de Gruyter, Berlin 2005, chapter 45.
  36. ^ Hans J. Holm: The New Arboretum of Indo-European "Trees". Can New Algorithms Reveal the Phylogeny and Even Prehistory of IE? In: Journal of Quantitative Linguistics. 14 (2), 2007, pp. 167-214.
  37. ^ Bernard C. Comrie: Language Universals and Linguistic Typology. Syntax and Morphology. University of Chicago Press, Chicago 1989, ISBN 0-226-11433-3 . (English)
  38. ^ Trask, 8.8
  39. Stone Age Language in Video Games ,, accessed on June 22, 2016

The information in this article comes mainly from the works by Fortson, Hoffmann, Jasanoff, Kloekhorst, Lehmann, Mailhammer, Meier-Brügger, Rings, Rix, Seebold and Sihler mentioned under literature .

This article was added to the list of excellent articles on September 3, 2008 in this version .