Proto-Indo-European language

Historical and geographical setting

Scholars have proposed multiple hypotheses about when, where, and by whom PIE was spoken. The Kurgan hypothesis, first put forward in 1956 by Marija Gimbutas, has become the most popular. It proposes that this proto-language was spoken by the Yamnaya culture in the general area of the Pontic steppe north of the Black Sea c. 3400 BCE.

Other theories include the Anatolian hypothesis, which posits that PIE spread out from Anatolia with agriculture beginning c. 7500–6000 BCE, the Armenian hypothesis, the Paleolithic continuity paradigm, and the indigenous Aryans theory. The last two of these theories are not regarded as credible within academia. Out of all the theories for a PIE homeland, the Kurgan and Anatolian hypotheses are the ones most widely accepted, and also the ones most debated against each other. Following the publication of several studies on ancient DNA in 2015, Colin Renfrew, the original author and proponent of the Anatolian hypothesis, has accepted the reality of migrations of populations speaking one or several Indo-European languages from the Pontic steppe towards Northwestern Europe.

Descendants

The antiquity of the earliest attestation (in units of 500 years) of each Indo-European group is: 2000–1500 BCE for Anatolian; 1500–1000 BCE for Indo-Aryan and Greek; 1000–500 BCE for Iranic, Celtic, Italic, Phrygian, Illyric, Messapic, South Picene, and Venetic; 500–1 BCE for Thracian and Ancient Macedonian; 1–500 CE for Germanic, Armenian, Lusitanian, and Tocharian; 500–1000 CE for Slavic; 1500–2000 CE for Albanian and Baltic.

The table lists the main Indo-European language families, comprising the languages descended from Proto-Indo-European.

Commonly proposed subgroups of Indo-European languages include Italo-Celtic, Graeco-Aryan, Graeco-Armenian, Graeco-Phrygian, Daco-Thracian, and Thraco-Illyrian.

There are numerous lexical similarities between the Proto-Indo-European and Proto-Kartvelian languages due to early language contact, as well as some morphological similarities—notably the Indo-European ablaut, which is remarkably similar to the root ablaut system reconstructible for Proto-Kartvelian.

Marginally attested languages

The Lusitanian language was a marginally attested language spoken in areas near the border between present-day Portugal and Spain. The Venetic and Liburnian languages known from the North Adriatic region are sometimes classified as Italic.

Albanian and Greek are the only surviving Indo-European descendants of a Paleo-Balkan language area, named for their occurrence in or in the vicinity of the Balkan Peninsula. Most of the other languages of this area—including Illyrian, Thracian, and Dacian—do not appear to be members of any other subfamilies of PIE, but are so poorly attested that proper classification of them is not possible. Forming an exception, Phrygian is sufficiently well-attested to allow proposals of a particularly close affiliation with Greek, and a Graeco-Phrygian branch of Indo-European is becoming increasingly accepted.

Phonology

Proto-Indo-European phonology has been reconstructed in some detail. Notable features of the most widely accepted (but not uncontroversial) reconstruction include:

• three series of stop consonants reconstructed as voiceless, voiced, and breathy voiced;
• sonorant consonants that could be used syllabically;
• three so-called laryngeal consonants, whose exact pronunciation is not well-established but which are believed to have existed in part based on their detectable effects on adjacent sounds;
• the fricative
• a vowel system in which and were the most frequently occurring vowels. The existence of as a separate phoneme is debated.

Notation

Vowels

The vowels in commonly used notation are:

Consonants

The corresponding consonants in commonly used notation are:

All sonorants (i.e. nasals, liquids and semivowels) can appear in syllabic position (commonly indicated by an underring for nasals and liquids). The syllabic allophones of *y and *w are realized as the surface vowels *i and *u respectively.

is an allophone of *s when next to a voiced consonant in certain positions.

is an allophone of *n before velar consonants.

Accent

The Proto-Indo-European accent is reconstructed today as having had variable lexical stress, which could appear on any syllable and whose position often varied among different members of a paradigm (e.g. between singular and plural of a verbal paradigm). Stressed syllables received a higher pitch and it is often said that PIE had a pitch accent. The location of the stress is associated with ablaut variations, especially between full-grade vowels ( and ) and zero-grade (i.e. lack of a vowel), but not entirely predictable from it.

The accent is best preserved in Vedic Sanskrit and (in the case of nouns) Ancient Greek, and indirectly attested in a number of phenomena in other IE languages, such as Verner's Law in the Germanic branch. Sources for Indo-European accentuation are also the Balto-Slavic accentual system and plene spelling in Hittite cuneiform. To account for mismatches between the accent of Vedic Sanskrit and Ancient Greek, as well as a few other phenomena, a few historical linguists prefer to reconstruct PIE as a tone language where each morpheme had an inherent tone; the sequence of tones in a word then evolved, according to that hypothesis, into the placement of lexical stress in different ways in different IE branches.

Morphology

Proto-Indo-European, like its earliest attested descendants, was a highly inflected, fusional language. Suffixation and ablaut were the main methods of marking inflection, both for nominals and verbs. The subject of a sentence was in the nominative case and agreed in number and person with the verb, which was additionally marked for voice, tense, aspect, and mood.

Root

Proto-Indo-European nominals and verbs were primarily composed of roots – affix-lacking morphemes that carried the core lexical meaning of a word. They were used to derive related words (cf. the English root "-friend-", from which are derived related words such as friendship, friendly, befriend, and newly coined words such as unfriend). As a rule, roots were monosyllabic, and had the structure (s)(C)CVC(C), where the symbols C stand for consonants, V stands for a variable vowel, and optional components are in parentheses. All roots ended in a consonant and, although less certain, they appear to have started with a consonant as well.

A root plus a suffix formed a word stem, and a word stem plus an inflectional ending formed a word. Proto-Indo-European was a fusional language, in which inflectional morphemes signaled the grammatical relationships between words. This dependence on inflectional morphemes means that roots in PIE, unlike those in English, were rarely used without affixes.

Ablaut

Many morphemes in Proto-Indo-European had short e as their inherent vowel; the Indo-European ablaut is the change of this short e to short o, long e (ē), long o (ō), or no vowel. The forms are referred to as the "ablaut grades" of the morpheme—the e-grade, o-grade, zero-grade (no vowel), etc. This variation in vowels occurred both within inflectional morphology (e.g., different grammatical forms of a noun or verb may have different vowels) and derivational morphology (e.g., a verb and an associated abstract verbal noun may have different vowels).

Categories that PIE distinguished through ablaut were often also identifiable by contrasting endings, but the loss of these endings in some later Indo-European languages has led them to use ablaut alone to identify grammatical categories, as in the Modern English words sing, sang, sung.

Noun

Proto-Indo-European nouns were probably declined for eight or nine cases:

• nominative: marks the subject of a verb. Words that follow a linking verb (copulative verb) and restate the subject of that verb also use the nominative case. The nominative is the dictionary form of the noun.
• accusative: used for the direct object of a transitive verb.
• genitive: marks a noun as modifying another noun.
• dative: used to indicate the indirect object of a transitive verb, such as Jacob in Maria gave Jacob a drink.
• instrumental: marks the instrument or means by, or with, which the subject achieves or accomplishes an action. It may be either a physical object or an abstract concept.
• ablative: used to express motion away from something.
• locative: expresses location, corresponding vaguely to the English prepositions in, on, at, and by.
• vocative: used for a word that identifies an addressee. A vocative is a noun of address where the identity of the party spoken to is set forth expressly within a sentence. For example, in the sentence, "I don't know, John", John is a noun of address, indicating the party being addressed.
• allative: used as a type of locative case that expresses movement towards something. It was preserved in Anatolian (particularly Old Hittite), and fossilized traces of it have been found in Greek. It is also present in Tocharian. Its PIE shape is uncertain, with candidates including *-h₂(e), *-(e)h₂, or *-a.

Late Proto-Indo-European had three grammatical genders:

• masculine
• feminine
• neuter

This system is probably derived from an older two-gender system, attested in Anatolian languages: common (or animate) and neuter (or inanimate) gender. The feminine gender only arose in the later period of the language. Neuter nouns collapsed the nominative, vocative and accusative into a single form, the plural of which used a special collective suffix *-h₂ (manifested in most descendants as -a). This same collective suffix in extended forms *-eh₂ and *-ih₂ (respectively on thematic and athematic nouns, becoming -ā and -ī in the early daughter languages) became used to form feminine nouns from masculines.

All nominals distinguished three numbers:

• singular
• dual
• plural

These numbers were also distinguished in verbs (see below), requiring agreement with their subject nominal.

Pronoun

Proto-Indo-European pronouns are difficult to reconstruct, owing to their variety in later languages. PIE had personal pronouns in the first and second grammatical person, but not the third person, where demonstrative pronouns were used instead. The personal pronouns had their own unique forms and endings, and some had two distinct stems; this is most obvious in the first person singular where the two stems are still preserved in English I and me. There were also two varieties for the accusative, genitive and dative cases, a stressed and an enclitic form.

Verb

Proto-Indo-European verbs, like the nouns, exhibited an ablaut system.

The most basic categorisation for the reconstructed Indo-European verb is grammatical aspect. Verbs are classed as:

• stative: verbs that depict a state of being
• imperfective: verbs depicting ongoing, habitual or repeated action
• perfective: verbs depicting a completed action or actions viewed as an entire process.

Verbs have at least four grammatical moods:

• indicative: indicates that something is a statement of fact; in other words, to express what the speaker considers to be a known state of affairs, as in declarative sentences.
• imperative: forms commands or requests, including the giving of prohibition or permission, or any other kind of advice or exhortation.
• subjunctive: used to express various states of unreality such as wish, emotion, possibility, judgment, opinion, obligation, or action that has not yet occurred
• optative: indicates a wish or hope. It is similar to the cohortative mood and is closely related to the subjunctive mood.

Verbs had two grammatical voices:

• active: used in a clause whose subject expresses the main verb's agent.
• mediopassive: for the middle voice and the passive voice.

Verbs had three grammatical persons: first, second and third.

Verbs had three grammatical numbers:

• singular
• dual: referring to precisely two of the entities (objects or persons) identified by the noun or pronoun.
• plural: a number other than singular or dual.

Verbs were probably marked by a highly developed system of participles, one for each combination of tense and voice, and an assorted array of verbal nouns and adjectival formations.

The following table shows a possible reconstruction of the PIE verb endings from Sihler, which largely represents the current consensus among Indo-Europeanists.

Numbers

Proto-Indo-European numerals are generally reconstructed as follows:

Rather than specifically 100, *ḱm̥tóm may originally have meant "a large number".

Particle

Proto-Indo-European particles were probably used both as adverbs and as postpositions. These postpositions became prepositions in most daughter languages.

Reconstructed particles include for example, *upo "under, below"; the negators *ne, *mē; the conjunctions *kʷe "and", *wē "or" and others; and an interjection, *wai!, expressing woe or agony.

Derivational morphology

Proto-Indo-European employed various means of deriving words from other words, or directly from verb roots.

Internal derivation

Internal derivation was a process that derived new words through changes in accent and ablaut alone. It was not as productive as external (affixing) derivation, but is firmly established by the evidence of various later languages.

Possessive adjectives

Possessive or associated adjectives were probably created from nouns through internal derivation. Such words could be used directly as adjectives, or they could be turned back into a noun without any change in morphology, indicating someone or something characterised by the adjective. They were probably also used as the second elements in compounds. If the first element was a noun, this created an adjective that resembled a present participle in meaning, e.g. "having much rice" or "cutting trees". When turned back into nouns, such compounds were Bahuvrihis or semantically resembled agent nouns.

In thematic stems, creating a possessive adjective seems to have involved shifting the accent one syllable to the right, for example:

• *tómh₁-o-s "slice" (Ancient Greek τόμος tómos) > *tomh₁-ó-s "cutting" (i.e. "making slices"; Greek τομός tomós) > *dr-u-tomh₁-ó-s "cutting trees" (Greek δρυτόμος drutómos "woodcutter" with irregular accent).
• *wólh₁-o-s "wish" (Vedic Sanskrit वर॑ vára) > *wolh₁-ó-s "having wishes" (Sanskrit व॒र vará "suitor").

In athematic stems, there was a change in the accent/ablaut class. The reconstructed four classes followed an ordering in which a derivation would shift the class one to the right:

acrostatic → proterokinetic → hysterokinetic → amphikinetic

The reason for this particular ordering of the classes in derivation is not known. Some examples:

• Acrostatic *krót-u-s ~ *krét-u-s "strength" (Sanskrit क्रतु॑ krátu) > proterokinetic *krét-u-s ~ *kr̥t-éw-s "having strength, strong" (Greek κρατύς kratús).
• Hysterokinetic *ph₂-tḗr ~ *ph₂-tr-és "father" (Greek πατήρ patḗr) > amphikinetic *h₁su-péh₂-tōr ~ *h₁su-ph₂-tr-és "having a good father" (Greek εὐπάτωρ eupátōr).

Vṛddhi

A vṛddhi derivation, named after the Sanskrit grammatical term, signifying "of, belonging to, descended from". It was characterised by "upgrading" the root grade, from zero to full (e) or from full to lengthened (ē). When upgrading from zero to full grade, the vowel could sometimes be inserted in an unexpected location, creating a different stem from the original full grade.

Examples:

• full grade *swéḱuro-s "father-in-law" (Vedic Sanskrit श्वशु॑र śváśura-) > lengthened grade *swēḱuró-s "relating to one's father-in-law" (Sanskrit श्वाशुर śvāśura, Old High German swāgur "brother-in-law").
• full grade *dyḗw-s > zero grade *diw-és "sky" (Sanskrit द्यौस् dyáus) > new full grade *deyw-o-s "god, sky god" (Sanskrit दे॒वस् devás, Ancient Greek Ζεύς Zeús, Latin deus, etc.). Note the difference in vowel placement, *dyew- in the full-grade stem of the original noun, but *deyw- in the vṛddhi derivative.

Nominalization

Adjectives with accent on the thematic vowel could be turned into nouns by moving the accent back onto the root. A zero grade root could remain so, or be "upgraded" to full grade like in a vṛddhi derivative. Some examples:

• PIE *ǵn̥h₁-tó-s "born" (Vedic Sanskrit जा॒त jātá) > *ǵénh₁-to- "child", literally "thing that is born"; Sanskrit जात jāta; German Kind).
• Ancient Greek λευκός leukós "white" > λεῦκος leûkos 'a kind of fish', literally "white one".
• Sanskrit कृ॒ष्ण kṛṣṇá 'dark' > कृष्ण॑स् kṛ́ṣṇas 'antelope' (literally "dark one").

This kind of derivation is likely related to the possessive adjectives, and can be seen as essentially the reverse of it.

Affixal derivation

Alternative morphological reconstructions

Roland A. Pooth has proposed an alternative, markedly different reconstruction of Proto-Indo-European morphology known as the "templatic model". In this analysis, PIE roots are interpreted as consonantal roots onto which transfixes (vowel melody templates) are superimposed (essentially replacing the ablaut of the standard reconstruction). Furthermore, the model proposes that PIE possessed a direct-inverse alignment, with verbal morphology distinguishing direct, inverse, and intransitive (or undirected) forms according to the direction of transitivity between participants.

Pooth also applied his model to a full reconstruction of "The North Wind and the Sun" in Proto-Indo-European.

Syntax

The syntax of the older Indo-European languages has been studied in earnest since at least the late nineteenth century, by such scholars as Hermann Hirt and Berthold Delbrück. In the second half of the twentieth century, interest in the topic increased and led to reconstructions of Proto-Indo-European syntax.

Since all the early attested IE languages were inflectional, PIE is thought to have relied primarily on morphological markers, rather than word order, to signal syntactic relationships within sentences. Still, a default (unmarked) word order is thought to have existed in PIE. In 1892, Jacob Wackernagel reconstructed PIE's word order as subject–verb–object (SVO), based on evidence in Vedic Sanskrit.

Winfred P. Lehmann (1974), on the other hand, reconstructs PIE as a subject–object–verb (SOV) language. He posits that the presence of person marking in PIE verbs motivated a shift from OV to VO order in later dialects. Many of the descendant languages have VO order: modern Greek, Romance and Albanian prefer SVO, Insular Celtic has VSO as the default order, and even the Anatolian languages show some signs of this word order shift. Tocharian and Indo-Iranian, meanwhile, retained the conservative OV order. Lehmann attributes the context-dependent order preferences in Baltic, Slavic and Germanic to outside influences. Donald Ringe (2006), however, attributes these to internal developments instead.

Paul Friedrich (1975) disagrees with Lehmann's analysis. He reconstructs PIE with the following syntax:

• basic SVO word order
• adjectives before nouns
• head nouns before genitives
• prepositions rather than postpositions
• no dominant order in comparative constructions
• main clauses before relative clauses

Friedrich notes that even among those Indo-European languages with basic OV word order, none of them are rigidly OV. He also notes that these non-rigid OV languages mainly occur in parts of the IE area that overlap with OV languages from other families (such as Uralic and Dravidian), whereas VO is predominant in the central parts of the IE area. For these reasons, among others, he argues for a VO common ancestor.

Hans Henrich Hock (2015) reports that the SVO hypothesis still has some adherents, but the "broad consensus" among PIE scholars is that PIE would have been an SOV language. The SOV default word order with other orders used to express emphasis (e.g., verb–subject–object to emphasise the verb) is attested in Old Indo-Aryan, Old Iranian, Old Latin and Hittite, while traces of it can be found in the enclitic personal pronouns of the Tocharian languages.