10. Sentence Syntax

10.1. The Sentence

A sentence is a form of words which contains a state­ment, a question, an exclamation, or a command.

The fundamental order of sentences in PIE appears to be OV. Support for this assumption is evident in the oldest texts of the materials attested earliest in the IE dialects. Examples of this include the following, from Fortson (2004): Hitt. nu=za MUŠilluyankaš DIM-an tarata ‘And the serpent overcame the Stormgod’; Ved. maruto ha enam na ajahu ‘Indeed the Maruts did not abandon him’; Lat. Eumolpus tanquam litterārum stūdiōsus utīque ātrāmentum habet ‘Eumolpus, so interested in learning, surely has (some) ink’; Runic ek hlewagastiz holtijaz horna tawido ‘I, Hlewagastiz of Holt, made (this) horn’; Toch. A. kāsu ñom-klyu tsraiśśi śäk kälymentwa sätkatar ‘Good fame of the strong spreads out in ten directions’.

NOTE. Lehmann (1974): “The fundamental order of sentences in these early dialects cannot be determined solely by frequency of sentence patterns. For, like other linguistic constructions, sentence patterns manifest marked as well as unmarked order. Marked order is expected in literary materials. The documents surviving from the earliest dialects are virtually all in verse or in literary forms of prose. Accordingly many of the individual sentences do not have the unmarked order, with verb final. For this reason conclusions about the characteristic word order of PIE and the early dialects will be based in part on those syntactic patterns that are rarely modified for literary and rhetorical effect: comparative constructions, the presence of postpositions and prepositions, and the absence of prefixes (…).”

Lehmann is criticised by Friedrich (1975) who, like Watkins (1976) and Miller (1975), support a VO prehistoric situation, probably SVO (like those found in ‘central’ IE areas), with non-consistent dialectal SOV findings. In any case (viz. Lehmann and Miller), an older PIH OV (VSO for Miller), as attested in Hittite, would have been substituted by a newer VO (SOV for Miller, later SVO through a process of verb transposition) – thus, all Indo-European dialects attested have evolved (therefore probably from a common post-Proto-Indo-European trend) into a modern VO.

Fortson (2004): “Part of the problem with it is arriving at a clear definition of a verb-final language. In the strict sense, a verb-final language is one where the verb always comes at the end of each clause unless other factors intervene. The only well-known older PIE language that meets this criterion is Hittite. No matter what the genre, no matter how stylistically marked the text, the verb in Hittite is always clause-final, with one exception – when it is fronted to the beginning of the clause for emphasis or contrast (…). None of the other old IE languages behaves so rigidly (…); there is essentially no position in the clause (on the surface at least) where the verb cannot appear.

It is usually stated that in these languages, the pragmatically neutral order is SOV. This may, in fact, be true, at least of some of them (such as Latin); but with so many word-order permutations possible (and frequent), clearly they cannot be called “verb-final” in the same way as Hittite. There are any number of reasons, according to current theory, why a verb may or may not appear as the last word in its clause.”

Clackson (2007): “The investigation of PIE word order in terms of rigid SVO and SOV patterns was rightly criticised as a ‘pseudo-problem’ by Watkins in a much-cited article of 1976. It would be wrong, however, to think that because word order cannot be expressed in terms of strict SOV or SVO patterns it is somehow unimportant, or that it was free. If PIE were a ‘non-configurational’ language, with completely freeword order, we would still have to explain why the unmarked place of the verb is sentence-final in Hittite, Sanskrit and Latin, and why word comparisons which reflect original juxtapositions of noun and dependent genitive agree in showing the order genitive – head noun (…) Agreements such as these could lead to the conclusion that the unmarked order was SOV for the PIE sentence, and head-final for the PIE noun phrase. But variation from these patterns is widely attested, particularly in poetic or highly stylised texts, which make up a large part of our corpus of many early IE languages.”

10.1.1. Kinds of Sentences

1. PIE sentences were either nominal, i.e. formed by nouns, or verbal, if they included a verb.

2. Sentences are formed by a subject and a predicate. The subject of a sentence is the person or thing spoken of.  The predicate is that which is said of the subject.

a. The subject is usually a noun or pronoun, or some word or group of words used as a noun.  However, because the personal endings of verbs encode the subject in them, it is not grammatically necessary to use an overt personal pronominal subject in addition. When subject pronouns occur, it is usually said that its use is emphatic (Fortson 2004); as, Lat. sīcutī ego accēpīas I understand it’ (Sallus, Bellum Catilinae 6.1), where ‘I’ serves to contrast his own understanding with the opinion of others.

However, in nominal sentences, an overt pronominal subject is generally required for clarity; as, O.Pers. adam navama, ‘I (am) the ninth’.

The syntax of possessive swos, own, is treated by Fortson (2004): “Reflexive adjectives (and pronouns) refer back to the grammatical subject of a sentence. But the possessive “swo-” had broader usage, to judge by the daughter languages: it could refer back not to the grammatical subject, but to newly introduced discourse material or to an older topic that is returned to. As an example of the former, consider Ved. tráya índrasya sóma sutsa santu devásya své káye sutapvna ‘Let the three somas be pressed for the god Indra in the soma drinker’s own house’ (translation following Brent Vine; soma was an intoxicating sacred drink). Here the grammatical subject is tráya. .. sóma ‘the three somas’ and the possessive své refers to the soma-drinker, who is newly introduced. Similar behavior can be found in other older IE languages.”

b. The predicate of a sentence may be a verb (as the dog runs), or it may consist of some form of esmi, be, and a noun or adjective which describes or defines the subject (as It is good). Such a noun or adjective is called a predicate noun or adjective.

3.  In Proto-Indo-European, simple sentences may be composed of only one word, a noun or a verb; as, ‘God!’ or ‘(it) rains’.

NOTE 1. Nominal sentences of this type are usually interjections and vocatives. Verbal sentences of this type include imperatives (at least of 2nd sg.) and impersonal verbs, which had never a subject in the oldest dialects attested; as, for Eng. (it) rains, cf. Goth. rigneiþ, Lat. pluit, Gk. ὓει, Skt. várati. It is believed that when IE dialects became SVO in structure, so that a subject was required, the third singular anaphoric pronoun, corresponding to it, German es, French il, etc., was introduced as subject in such sentences. Such pronouns were introduced because SVO languages must have subjects in sentences, as do intransitive verbs in any OV language. Such verbs could be supplemented by substantives in various cases, among them the accusative. These constructions are especially prominent for verbs referring to the emotions; as, Skt. kitavátatāpa (lit. with-regard-to-the-gambler there-is-pain) ‘it pains the gambler’, Lat. miseret, pudet, taedet ‘It makes one pitiful, ashamed, bored’. Compare also Cicero’s Lat. eōrum nōs miseret (lit. ‘of-them us it-makes pitiful/there-is-pity’) ‘we feel pity for them’ or O.H.G. thes gánges thih nirthrúzzi (lit. ‘of-the way you may-there-not-be-weariness’) ‘Do not let yourself be wearied of the way’. In PIE sentences various case forms could be used with verbs. The simplest sentences may consist of verbs accompanied by nouns in seven of the eight cases; only the vocative is not so used. The nouns fill the role of objects or, possibly better stated, of complements (Lehmann 1974).

NOTE 2. Besides the simple sentence which consists only of a verb, a simple sentence in the early dialects and in PIE could consist of a verb accompanied by a noun or pronoun as complement. A subject however wasn’t mandatory. Nor were other constructions which may seem to be natural, such as indirect objects with verbs like ‘give’. The root *dō- or in its earlier form *deH- had in its simplest sense the meaning ‘present’ and was often unaccompanied by any nominal expression (Lehmann 1974).

10.1.2. Nominal Sentence

1. Nominal sentences, in which a substantive is equated with another substantive, an adjective, or a particle, make up one of the simplest type of sentence in PIE.

Such a type of sentence is found in almost every IE dialect; cf. (Lehmann 1974) Hitt. attaš aššuš ‘the father (is) good’, Skt. tvá várua, ‘you (are) Varuna’, O.Pers. adam Dārayavauš ‘I (am) Darius’, Gk. emoì d’ ákhos ‘and to me (there is) pain’, Lat. omnia praeclara raraall the best things (are) rare’. Apart from a noun or adjective, the predicate could be an adverb; as, Hitt. mŠippa-LÚ-iš=wa=kan ŪL anda ‘Sippazitis (is) not in (it)’; Gk. metà dè glaukō̃pis Athénē ‘and Athena of glaucous eyes (was) with (them)’.

In all dialects, however, such sentences were restricted in its use to a especially formal use or, on the contrary, they are found more often than originally in PIE. Thus, in Latin and Germanic dialects they are found in proverbs and sayings, as in Old Irish; in Greek it is also found in epic and poetry. In Balto-Slavic dialects the pure nominal sentence has become the usual type of nominal sentence, even when the predicate is an adverb or an adverbial case. However, such a use, which is more extended in modern dialects (like Russian) than in the older ones (as Old Slavic), is considered the result of Finno-Ugrian influence. (Lehmann 1974).

2. An action, state, or event could be syntactically backgrounded using a construction called an absolute. Typically the absolute consisted of a noun modified by a participle – semantically equivalent to a subject plus verb – in an oblique case (Fortson 2004).

Thus Latin has ablative absolutes (hīs rēbus gestīs lit. ‘these things having been done’, i.e. ‘after these things were done’ or ‘because these things were done’), Greek has genitive absolutes (Homeric aékontos emeĩowith me being unwilling’, pàr émoige kaì hálloi hoi ké mé timsousi, ‘near me (there are) others who [particle] will praise me’), Vedic Sanskrit has locative absolutes (ucchántyām uśásiwith dawn shining forth’, or havyaír agnír mánua īrayádhyai, ‘Agni must be prayed with the sacrifices of men’,), and Gothic and Old Church Slavonic have dative absolutes (Goth. imma rodjandin and O.C.S. jemu glagoljǫščemuwith him speaking, while he is/was speaking’).

NOTE. PIE surely had such constructions too, although which case or cases were used is debated.

3. In addition to such expansions by means of additional nouns in nonrequired cases, sentences could be expanded by means of particles.

NOTE. For Lehmann (1974), three subsets of particles came to be particularly important. One of these is the set of preverbs, such as ā. Another is the set of sentence connectives, such as nu. The third is the set of qualifier expressions, e.g., PIE /(must) not’. An additional subset, conjunctions introducing clauses, will be discussed below in the section on compound clauses. Preverbs are distinctively characterised by being closely associated with verbs and modifying their meaning. In their normal position they stand directly before verbs (Watkins 1964).

Generally, thus, concordance governed both members of the pure nominal sentence. Unlike the personal verb and its complements (governed by inflection), the nominal sentence showed a strong reliance on concordance between subject and predicate as a definitory feature: both needed the same case, and tended to have the same number and gender (Lehmann 1974).

The Copulative Verb

Nominal sentences are not always simple clauses without copula; examples are found with an explicit copula with stylistic and semantic change (Ramat 1993): Hitt. LÚ.ULÙ.LU=ku GUD=ku UDU=ku ēšzi ‘whether he is man or ox or sheep’.

The copulative verb esmi is only necessary when introducing late categories in the verbal morphology, like time and mood. Therefore, when the mood is the indicative, and the time is neuter (proverbs without timing, or present with semantic neuter) there is no need to use esmi.

NOTE. The basic form of nominal sentences has been a matter of dispute. Some Indo-Europeanists propose that the absence of a verb in nominal sentences is a result of ellipsis and assume an underlying verb esmi (Benveniste 1950). They support this assumption by pointing to the requirement of such a verb if the nominal sentence is in the past tense; cf. Hitt. ABU.I̯A genzuu̯alaš ešta ‘My father was merciful’. On the contrary, Meillet (1906-1908), followed by Lehmann (1974) and Mendoza (1998), consider that nominal sentences did not require a verb but that a verb might be included for emphasis. This conclusion may be supported by noting that the qualifiers which were found in PIE could be used in nominal sentences without a verb. As an example we may cite a Hittite sentence which is negative and imperative, 1-aš 1-edani menahhanda idāluš ‘One should not be evil toward another one’. Yet, if a passage was to be explicit, a form of esmi could be used, as in Skt. nákir indra tvád úttaro jy asti “No one is higher than you, Indra, nor greater’.

Fritz (in Meier-Brügger 2003): “Nominal phrases are not simply verbal phrases without a finite verb (with what is called ellipsis of the copula), but rather constitute an independent type of clause. Thus, the predicate noun in nominal phrases is always stressed, unlike the verbal predicate in verbal phrases. In fact the term ‘elipsis’ is not exact, since the copula is not essential. Contrarily, the use of the copula should rather be seen as an adaptation to the common pattern of verbal phrases, which always feature a finite verb form. This use of the copula is in fact a sort of explicative signification, in which the content of the copula is expressed through the connection of the various sentence elements and is given particular emphasis alone through an independent linguistic symbol of comparable meaning.”

On the original meaning of esmi, since Brugmann (1925) it is reconstructed as ‘exist’ hence its use as a copulative verb through constructions in which the predicate express the existence of the subject, as in Hom. Gk. eím’ Odusseús Laertiádēs, hós… ‘I am Odisseus, son of Laertes, the one who…’ (Mendoza 1998).

10.1.3. Verbal Sentence

The simplest structure of the common Indo-European sentence consists of a verb, i.e. the carrying out of an action. In it, none of the verbal actors (subject and object) must be expressed – the subject is usually not obligatory, and the object appears only when it is linked to the lexical nature of the verb.

NOTE. The oldest morphological categories, even time, were expressed in PIE through lexical means, and remains are found of such a system; cf. Hitt. -za (reflexive), modal particles in Gk. and O.Ind., modal negation in some IE dialects, or the simple change in intonation, which made interrogative or imperative a declarative sentence – in fact, the imperative lacks a mark of its own.

The relationship between the subject and the object is expressed through the case.

There is no clear morphological distinction between transitive and intransitive verbs in Proto-Indo-European.

NOTE. Some Indo-European dialects have specialised certain verbal suffixes as transitives (causatives) or intransitives, as Gk. -en, Gmc. -io, Lat. -a, etc., while in some others a preverb combined with a verbal root makes the basic verb transitive or intransitive.

When subjects are explicitly expressed, the nominative is the case employed.

NOTE. Expression of the subject is the most prominent extension of simple sentences to include more than one substantival expression. Apart from such explicit mention of the subject, predicates may consist of verbs accompanied by two or more nouns, in cases which supplement the meanings of the verbs (see below). Such constructions must be distinguished from the inclusion of additional nouns whose case forms indicate adverbial use.

Predicates with two or more substantives

Few verbs are mandatorily accompanied by two nouns (Lehmann 1974):

1. The use of the dative in addition to the accusative. This is made almost obligatory with verb dō-, give; as in Skt. tbhiām ena pári dehi (lit. those-two-DAT. him-ACC. over you-give), ‘Give him over to those two’; Gk. dō̑ra phérein tō̑i patrí ‘give presents to the father’ (Mendoza 1998).

2. The instrumental and ablative, as Skt. áhan vr̥trám... índro vájrea ‘Indra killed ... Vṛtra with his bolt’; Skt. tvá dásyūm̐r ókaso agna āja ‘You drove the enemies from the house, O Agni’.

NOTE.  While the addition to these sentences which is indicated by the nouns in the instrumental and the ablative is essential for the meaning of the lines in their context, it does not need to be included in the sentence for syntactic reasons.

3. The causative accompanied by two accusatives, as Skt. devn̐ uśatapāyayā haví (lit. gods desiring you-cause-to-drink libation) ‘Make the desiring gods drink the libation’.

In such sentences the agent-accusative represents the object of the causative element: as Arthur A. Macdonell indicated (1916), in a corresponding simple sentence this noun would have been given in the nominative, as Skt. dev haví pibantiThe gods drink the libation’.

Accordingly a simple verb in PIE was at the most accompanied by one substantive, unless the additional substantive was complementary or adverbial.

10.2. Sentence Modifiers

10.2.1. Intonation Patterns

1. According to Fritz (in Meier-Brügger 2003), with regard to sentence accent, one may note that the word that begins the sentence is stressed. Sentence-initial position implies the function of establishing the topic: In nominative language, the subject in sentence-initial position is considered the normal, unmarked type. In interrogative sentences, the nominal element, about which the question is asked, establishes the theme. It is thus the interrogative pronoun that begins the sentence

NOTE. An enclitic is added as a second word in the sentence; a further enclitic is added as a third word. This is known as the (Delbrück-)Wackernagel Law, see §10.6.4.

2. The sentence was characterised in PIE by patterns of order and by selection (Lehmann 1974):

A. Selection classes were determined in part by inflection, in part by lexical categories, most of which were covert.

NOTE. Some lexical categories were characterised at least in part by formal features, such as abstract nouns marked by -ti-, nouns in the religious sphere marked by -u- and collectives marked by *-h.

B. In addition to characterisation by means of order and categories of selection, the sentence was also delimited by intonation based on variations in pitch.

2. To the extent that the pitch phonemes of PIE have been determined, a high pitch may be posited, which could stand on one syllable per word, and a low pitch, which was not so restricted.

Words were characterised on one syllable by a high pitch accent, unless they were enclitic, that is, unmarked for accent.

3. Accented words could lose their high pitch accent if they were placed at specific positions in sentences:

A. Vocatives lost their accent if they were medial in a sentence or clause; and finite verbs lost their accent unless they stood initially in an independent clause or in any position in a dependent clause in Vedic. These same rules may be assumed for PIE. On the basis of the two characteristic patterns of loss of accent for verbs, characteristic patterns of intonation may also be posited for the IE sentence.

B. Judging on the basis of loss of high pitch accent of verbs in them, independent clauses were characterised by final dropping in pitch. For in unmarked order the verb stands finally in the clause; as, purodhm evá gacchati (lit. priesthood verily he-attains) ‘He attains the priesthood’.

C. In marked order on the other hand it stands initially. H. S. Ananthanarayana investigated the accent patterns in accented Vedic texts, particularly in the Taittirīya Brāhmaa, and concluded on the basis of the interpretation of sentences with similar lexical material that sentences with initial verb are marked.

 Thus, in contrast with the previous example, the following indicates “emphasis” of the verb (Ananthanarayana 1970): gácchati pratiṣṭhmHe attains stability’  Since gacchati in the other example has no high pitch accent, and since other such sentences have a similar distribution of accents, it may be concluded that sentences with normal, unmarked meaning have a final lowered pitch accent. This might be indicated with #.

Clauses, however, which are marked either to convey emphasis or to indicate subordination, do not undergo such lowering. They may be distinguished with final || (Ananthanarayana 1970): yát stríyam upeyt || nírvīryas syāt # (lit. if woman he-may-approach impotent he-may-become) ‘If he were to approach a woman, he might become impotent’.

The intonation pattern indicated by # apparently conveyed the notion of a simple, nonemphatic utterance, whether a statement, question, or command: kásmai nú satrám āsmahe # (lit. why indeed sacrifice we-perform) ‘Why should we perform sacrifice?’

The intonation pattern indicated by || apparently conveyed the notion of an emotional or emphatic utterance or one requiring supplementation, as by another clause. These conclusions are supported by the patterns found in Germanic alliterative verse. For, as is well known, verbs were frequently placed by poets in the fourth, nonalliterating, metrically prominent position in the line: þeodcyninga þrym gefrūnon (lit. of-people’s-kings glory we-heard-of) ‘We heard of the glory of the kings of the people’.

This placing of verbs, retained by metrical convention in Germanic verse, presumably maintains evidence for the IE intonation pattern. For, by contrast, verbs could alliterate when they stood initially in clauses or in subordinate clauses; egsode eorlas, syððan ǣrest wearð (lit. he-terrified men since first he-was) ‘He terrified men from the time he first was [found]’; þenden wordum wēold wine Scyldinga (lit. as-long-as with-words he-ruled the-friend of-the-Scyldings) ‘As long as the friend of the Scyldings ruled with his words.’

The patterns of alliteration in the oldest Germanic verse accordingly support the conclusions that have been derived from Vedic accentuation regarding the intonation of the Indo-European sentence, as do patterns in other dialects.

D. Among such patterns is the preference for enclitics in second position in the sentence (Wackernagel 1892). Words found in this position are particles, pronouns, and verbs, which have no accent in Vedic texts. This observation of Wackernagel supports the conclusion that the intonation of the sentence was characterised by initial high pitch, with the voice trailing off at the end. For the enclitic elements were not placed initially, but rather they occupied positions in which unaccented portions of words were expected, as in Skt. prāvep bható mādayanti, ‘The dangling ones of the lofty tree gladden me’. The pronoun ‘me’, like other such enclitics, makes up a phrase with the initial word; in this way it is comparable to unaccented syllables of individual words, as in Skt. pravātej írie várvtānā (lit. in-windy-place on-dice-board rolling)‘[born] in a windy place, rolling on the dice-board’.

10.2.2. Sentence Delimiting Particles

1. The particles concerned are PIE nu, so, to, all of them introductory particles.

Their homonymity with the adverb nu, nun and the anaphoric pronoun was one of the reasons earlier Indo-Europeanists failed to recognise them and their function. Yet Delbrück had already noted the clause-introducing function of Skt. sa (1888), as in Skt. tásya tniśīri prá cicheda. (lit. his the heads off he-struck) yát somapnam sa tátakapíñjala sám abhavat (lit. Ptc. what soma-drinking it-was from-that hazel-hen [grouse] together it-became] ‘He struck off his heads. From the one that drank soma, the hazel-hen was created’. Delbrück identified sa in this and other sentences as a particle and not a pronoun, for it did not agree in gender with a noun in the sentence. But it remained for Hittite to clarify the situation.

In Hittite texts the introductory use of the particles is unmistakable (J.Friedrich 1960); ta and šu occur primarily in the early texts, nu in the later, as illustrated in the following Old Hittite example (Otten and Souček 1969): šer-a-ššan GAD-an pešiemi šu-  LÚ-aš natta aušzi (over-and-Ptc. cloth I-throw Ptc.-  them man not sees) ‘I throw a cloth over it and no one will see them’ (Lehmann 1974).

2. Besides such an introductory function (here as often elsewhere translated ‘and’), these particles were used as first element in a chain of enclitics, as in n-at-ši ‘and it to-him’, nu-mu-za-kan ‘and to-me self within’ and so on.

In Homeric Greek such strings of particles follow different orders, but reflect the IE construction, as in: oudé nu soí per entrépetai phílon êtor, Olúmpie (lit. not-indeed and to-you but it-turns ‘dear’ heart Olympian) ‘But your heart doesn’t notice, Zeus’. As the translation of per here indicates, some particles were used to indicate the relationships between clauses marking the simple sentence (Lehmann 1974).

3. Many simple sentences in PIE would then be similar to those in Hittite and Vedic Sanskrit. Among the simplest is Skt. tám índro didvea ‘Indra hated him’. Presumably tam is a conflated form of the particle ta and the enclitic accusative singular pronoun; the combination is attested in Hittite as ta-an (J. Friedrich 1960). Similar examples from the other early dialects could be cited, such as the Italic inscription of Praeneste, or the Germanic Gallehus inscription: Ek HlewagastiR HoltijaR horna tawido, ‘I, Hlewagastir of Holt, made the horn’. In these late texts, the subject was mandatory, and accordingly two nominal forms had come to be standard for the sentence. If however the subject is not taken into consideration, many sentences contained only one nominal element with verbs, in the early dialects as well as in PIE (Lehmann 1974).

10.3. Verbal Modifiers

10.3.1. Declarative Sentences

1. The injunctive has long been identified as a form unmarked for mood and marked only for stem and person. It may thus be compared with the simplest form of OV languages.

 By contrast the present indicative indicates “mood”. We associate this additional feature with the suffix -i, and assume for it declarative meaning.

2. As Lehmann (1974) says, “Yet it is also clear that, by the time of Vedic Sanskrit and, we assume, Late PIE, the injunctive no longer contrasted directly with the present indicative. We must therefore conclude that the declarative qualifier was expressed by other means in the sentence. We assume that the means of expression was an intonation pattern. For, in normal unmarked simple sentences, finite unaccented verbs stood finally in their clause, as did the predicative elements of nominal sentences; Delbrück’s repeatedly used example may be cited once again to illustrate the typical pattern: víśa katríyāya balí haranti ‘The villagers pay tribute to the prince’. Since the verb haranti was unaccented, i.e., had no high pitch, we may posit for the normal sentence an intonation pattern in which the final elements in the sentence were accompanied by low pitch.”

Lehmann (1974) supports this assumption by noting that a distinctive suprasegmental was used in Vedic to distinguish a contrasting feature, interrogation or request (Wackernagel 1896). This marker, called pluti by native grammarians, consisted of extra length, as in ágnā3i ‘O fire’ (3 indicates extra length). But a more direct contrast with the intonation of simple sentences may be exemplified by the accentuation of subordinate clauses. These have accented verbs, as in the following line from the Rigveda: antáś ca pr áditir bhavāsi ‘If you have entered inside, you will be Aditi’. As the pitch accent on ágā indicates, verbs in subordinate clauses maintained high pitch, in contrast with verbs of independent clauses like bhavāsi. We may conclude that this high pitch was an element in an intonation pattern which indicated incompleteness, somewhat like the pattern of contemporary English.

3. Evidence from other dialects supports the conclusion that, in late PIE, declarative sentences were indicated by means of an intonation pattern with a drop in accentuation at the end of the clause.

NOTE. In Germanic verse, verbs of unmarked declarative sentences tend to occupy unaccented positions in the line, notably the final position (Lehmann 1956). Although the surface expression of accentuation patterns in Germanic is stress, rather than the pitch of Vedic and PIE, the coincidence of accentuation pattern supports our conclusions concerning PIE intonation.

10.3.2. Interrogative Sentences

1. The interrogation was apparently also indicated by means of intonation, for some questions in our early texts have no surface segmental indication distinguishing them from statements, for example, Plautus Aulularia 213, aetatem meam scis, ‘Do you know my age?’.

NOTE. Only the context indicates to us that this utterance was a question; we may assume that the spoken form included means of expressing interrogation, and in view of expressions in the later dialects we can only conclude that these means were an intonation pattern.

2. Questions are generally classified in two groups:

·  Those framed to obtain confirmation, yes/no questions (Bestätigungsfragen). This feature accompanies statements in which a speaker sets out to elicit information from the hearer.

·  Those framed to obtain clarification (Verdeutlichungsfragen).

3. Yes/no questions (Bestätigungsfragen) were made (Clackson 2007):

 a) By an intonation pattern alone, as noted above. That is the reconstruction favoured by most of those who have addressed the issue (Delbrück 1893-1900, Meier-Brügger 2003), because of the finds in Hittite and Vedic Sanskrit. It might therefore be considered one of the oldest means to express interrogation of any type, including yes/no questions. This is most probably the older situation in PIE.

b) Disjunctive questions can be formed by juxtaposition of a verb with a negated verb, as in the following example of Vedic Sanskrit prose text (3 indicates extra length): chinátti s ná chinattī3 [cuts she not cuts pluti] ‘Does she divide or not?’..

NOTE. For Clackson (2007): “These different reconstructed hypotheses are not mutually exclusive: it is possible that different types of ‘yes-no’ question formation existed alongside each other in PIE. Indeed, systems of marking through intonation exist alongside other systems in many languages of the world. In French, for example, there are three different ways of forming ‘yes-no’ questions: Il vient?, Est-ce qu’il vient? and Vient-il? all mean ‘Is he coming?’.” 

c) By an interrogative affix or particle. Such means of expression for interrogation are found in most IE languages, apparently from a late development, since the particles used are different. Two of them have been reconstructed for PIE, though:

i. The particle nu, found in Greek and Vedic Sanskrit as interrogative particle, is also extremely widely used in non-interrogative sentences in Hittite, as well as in Sanskrit and Greek. It was probably then mainly a S.LIE resource.

ii. Lehmann (1974), following Delbrück (1893-1900) and Eichner (1971) argues that Lat. -ne, was the original interrogative particle, since its post-placement accorded with the typology of OV languages, in which interrogative particles are placed sentence finally. According to Minton Warren, it “occurs about 1100 times in Plautus and over 40 times in Terence” (1881). Besides expressions like Lat. egone ‘Me?’, sentences like the following occur (Plautus Asinaria 884): Aúdin quid ait? Artemona: Aúdio. ‘Did you hear what he is saying? Artemona: yes’.

Other evidence for a postponed particle for expressing interrogation is found in Avestan, in which -na is suffixed to some interrogatives, as in Av. kas- ‘who (then)?’; and in Germanic, where na is found finally in some questions in Old High German. Old Church Slavic is more consistent in the use of such a particle than are these dialects, as in chošteši li ‘Do you wish to?’ This particle is also used in contemporary Russian.

The particle used to express interrogation in Latin, Avestan, and Germanic is homophonous with the particle for expressing negation, PIE ne.

NOTE. It is not unlikely that LIE an of questions is behind same ne/ particle used for the negative. As the interrogative particle, however, it has been lost in most dialects. After Lehmann (1974), its loss is one of the indications that late PIE was not a consistent OV language. After Mendoza, the fact that such interrogatives of a yes/no-answer are introduced by different particles in the oldest attested dialects means that no single particle was generalised by Late Indo-European; cf. Goth. niu, Lat. -ne, nonne, num, Gk. ἣ, νὐ , Skt. nu, Sla. li. However, the common findings of Hittite, Indo-Iranian, Germanic and Latin are similar, if not the same.

4. The partial interrogative sentences are those which expect an aclaratory answer (Verdeutlichungsfragen), equivalent to English ‘What…?’, ‘Where…?’, ‘Who…?’, and so on. They are introduced in PIE by pronominal or adverbial forms derived from interrogative qi-/qo-, always placed initially but for marked sentences, where a change in position is admitted to emphasise it.

NOTE. In some languages, interrogatives may be strengthened by the addition of posposed particles with interrogative sense, as in Av. kaš-na (<*kwos-ne). Such forms introduce indirect interrogatives when they ask about a part of the sentence. Indirect interrogatives in the form of total interrogatives (i.e., not of yes/no-answer) are introduced by particles derived from direct interrogative particles (when there are) or by conditional conjunctions; as Hitt. man.

5. According to Clackson (2007), question words are typically fronted and  followed by enclitics, as Vedic kás te mātára vidhávām acakrat (lit. who you-gen mother-acc widow-acc he-made) ‘Who made your mother a widow?’; Vedic kásya bráhmāni jujusur yúvāna lit. ‘of whom the formulas like the-young-men?’;

Where an element precedes the question word, it makes sense to interpret this as a left-detached element, and it appears that in Vedic Sanskrit, as in Hittite, left-detachment does not count in the calculation of second position; as Vedic índra kím asya sakhyé cakāra (lit. Indra-nom what-acc he-gen friendship-loc he-did), ‘As for Indra, what did he do in his friendship?’; Vedic brahmā vasaparyati ‘priest which you honors?’ See §10.6.3 for more on emphasis.

10.3.3. Negative Sentences

1. Indications of negation, by which the speaker negates the verbal means of expression, commonly occupy third position in the hierarchy of sentence elements.

2. We can only posit the particles ne and mē/nē, neither of which is normally postposed after verbs.

NOTE. For prohibitive particle mē, compare Gk. m, O.Ind.,Av.,O.Pers. , Toch. mar/, Arm. mi, Alb. mos. In other IE dialects it appears as nē, cf. Goth. ne, Lat. nē (also as modal negation), O.Ir. ni. It is not clear whether Hitt. lē is ultimately derived from mē or nē, although Clackson (2007) reconstructs a common Anatolian *nē, due to Luv. ni(s), Lyc. ni. Although mē is sometimes reconstructed as the ‘original’ PIE particle, the Anatolian finds don’t let us decide which form is older. Apparently, S.LIE and Tocharian had mē, while Anatolian and North-West IE seem to have used nē.

PIE ne is found as Goth.,O.H.G. ni, Lat. - (e.g. in nequis) O.Ind. , O.Sla. ne, etc. Sometimes it is found in lengthened or strengthened forms as Hitt. natta, Lat. non, Skt. ned, etc. A common PIE lengthened form is nei, which appears in Lat. ni, Lith. neî, Sla. ni, etc.

3. The negative element ne was not used in compounding in PIE (Brugmann 1904); the privative prefix -, un-,had this function.

Comparative evidence suggests, following Fortson (2004), that certain classes of words were preferentially negated not with the adverb but with the privative prefix; among these words were participles and verbal adjectives. Greek and Latin, for example, ordinarily use their negative adverbs when negating participles, but some fixed archaic constructions point to an earlier time when the privative prefix was used instead, as Homeric Gk. a-ékontos emeĩo ‘with me being unwilling, against my will’, Lat. mē īn-sciente ‘with me not knowing’, in-uītus ‘unwilling’ (later replaced by nōn uolēns ‘not willing’), im-prūdēns ‘not knowing beforehand’ (later nōn prouidēns). Compare also Av. an-usat- ‘not wanting’, Goth. un-agands ‘not fearing’.

4. In the oldest languages, negation seems to have been preverbal; Vedic nákis, Gk. oú tis, mḗ tis, Lat. nēmo, OHG nioman ‘no one’, and so on. If the negation has scope over a single word or constituent, it usually directly precedes that constituent. Sentential negation typically directly precedes the verb, as in English. But it could also be moved toward the front of the sentence for emphasis. (Fortson 2004).

NOTE. Moreover, there is evidence for proposing that other particles were placed postverbally in PIE (Delbrück 1897). Delbrück has classified these in a special group, which he labels particles. They have been maintained postpositively primarily in frozen expressions: ē in Gk. egnē, ge in égōge ‘I’ (Schwyzer 1939). But they are also frequent in Vedic and early Greek; Delbrück (1897) discusses at length the use of Skt. gha, Gk. ge, and Skt. sma, Gk. mén, after pronouns, nouns, particles, and verbs, cf. Lat. nōlo < ne volo, Goth. nist< ni ist, and also, negative forms of the indefinite pronoun as O.Ind. m-kis, -kis, Lat. ne-quis, etc. which may indicate an old initial absolute position, which could be also supported by the development of correlative forms like Lat. neque, etc., which combine negation and coordination. Lehmann, on the contrary, believes in an older postposed order, characteristic of OV languages, because of the usually attributed value of emphasis to the initial position of negation, postverbal negation examples (even absolute final position in Hittite and Greek), the old existence of the form nei, as well as innovative forms like Lat. ne-quis or Gk. -tis (Lehmann 1974).

It is therefore safe to assume that in post-LIE times negation was usually preverbal, as in modern Romance languages (cf. Fr. n’est, Spa. no es, etc.), but it could be placed word-initially in emphatic contexts, and it is also found postponed in some archaic lexical or syntactic remains of the older IE languages, as it is found in modern Germanic languages (cf. Eng. is not, Ger. ist nicht, etc.).

5. Prohibitive sentences have a different negative particle, mē/nē. The older IE languages seem to have used the unmarked indicative (the so-called ‘injunctive’) for prohibitions, as some relic forms in Vedic might show. In Tocharian and Hittite, the tense-marked indicative was used. However, the tendency to replace the indicative with the imperative in prohibitions is seen in Greek and Latin, where the imperative became the marker of all commands, being mē/nē the modal negative.

For example, following Clackson (2007), while the ‘original’ situation would have been cemt, he came (aorist), ne cemt, he did not come (aorist), cemje, come, mē/nē cemjes, stop coming, this was reinterpreted and the imperative was used instead (hence a negative particle added to the positive utterance), i.e. mē/nē cemje, stop coming.

NOTE. Clackson (2007) continues: “Note also that several languages have opted to use modal forms (subjunctive and optative) in prohibitions. The most likely explanation for this is that speakers have extended the secondary functions of these modal forms, which include marking requests, wishes and other directive expressions, to embrace negative commands as well.”

According to Clackson (2007), some older IE languages show a difference between inhibitives, commands to stop doing something that the hearer is engaged in, and preventatives, commands or warnings not to do something in the future; that differentiation is found in Indo-Iranian, Tocharian, and apparently in Celtic too. Although the constructions differ, the Indo-Iranian differentiation could have been the original one: the present stem forms inhibitives, while the aorist stem forms preventatives.

10.4. Nominal Modifiers

10.4.1. Adjective and Genitive Constructions

1. Proto-Indo-European attributive adjectives were normally preposed.

Delbrück (1900) summarises the findings for Vedic, Greek, Latin, Lithuanian, and Germanic, giving examples like the following from Vedic: śvet párvatā ‘white mountains’. Lehmann (1974) adds an example of Hitt. šuppi watar ‘pure water’.

In marked constructions adjectives might be postposed, as in Ved. áśvaśvetá ‘a white horse, a gray’.

2. The position of the attributive genitive is the same as that of the attributive adjective. A striking example is given from the Old English legal language (Delbrück 1900): ōðres mannes hūses dura ‘the door of the house of the other man’.

3. Like the adjective construction, the attributive-genitive construction may have the modifier postposed for marked effect, as is sómasya in Skt. nas tátasyād íti? prathamabhaksá evá sómasyarjña íti (lit. what us then it-might-be Ptc. first-enjoyment Ptc. of-soma) ‘What might then happen for us? The first enjoyment of [Prince] Soma’. (Delbrück 1878)

NOTE 1. The relatively frequent marked use of the genitive may be the cause for the apparently free position of the genitive in Greek and Latin. The ambivalent order may also have resulted from the change of these languages toward a VO order. But, as Delbrück indicates, the preposed order is well attested in the majority of dialects. This order is also characteristic of Hittite (J. Friedrich 1960).

NOTE 2. In accordance with Lehmann’s (1974) views on syntactic structure, the attributive genitive, like the attributive adjective, must be derived from an embedded sentence. The sentence would have a noun phrase equivalent with that in the matrix sentence and would be a predicate nominal sentence. Such independent sentences are attested in the older dialects. Delbrück gives a number of examples, among them: Skt. aṣṭaú ha vaí putrádites (eight Ptc. Ptc. sons of-Aditi) ‘Aditi had eight sons’; Skt. áhar devnām sīt (lit. day of-gods it-was) ‘Day belonged to the gods’. These sentences accordingly illustrate that the genitive was used in predicate nominative sentences to convey what Calvert Watkins has labelled its primary syntactic function: the sense “of belonging”. When such a sentence was embedded in another with an equivalent NP, the NP was deleted, and the typical genitive construction resulted. Hittite also uses s as a genitive as well as a nominative marker. For “genitives” like haššannaššaš ‘(one) of his race’ can be further inflected, as in the accusative haššannaš-šan ‘(to one) of his race’ (J. Friedrich).

4. Fortson (2004): Common to all the older languages was the ability of nouns and their modifiers to be separated by intervening elements, yielding what are called discontinuous or distracted noun phrases (a construction called hyperbaton in Greek and Latin grammar): Cuneiform Luv. alati awienta Wilušati ‘they came from steep Wjlusa’ (KBo 4.11:46), Gk. ándra moi énnepe Moũsa polútropon ‘tell me, Muse, of the resourceful man(Odyssey 1.1), Lat. magnā cum laude ‘with greatpraise’, Old Irish Marta for slúaig saithiu ‘on the swarm of the host of March(Felire Oengusso, March 31). Distraction of other types of phrases was common as well. The technical details of distraction are not well understood; in some cases, it is the result of part of a phrase being moved to a position of emphasis or contrast.

10.4.2. Compounds

1. Nominal compositum or nominal composition is the process of putting two or more words together to form another word. The new word, called a compound word, is either a noun or an adjective, and it does not necessarily have the same meaning as its parts.

2. According to their meaning, compounds can be broadly classified in two types (Fortson 2004):

a. Endocentric (or determinative), as blackbird, ‘(a type of) black bird’, when the compound is essentially the sum of its parts, and its referent (a type of bird) is one of the compound members itself (usually the second one, as here). Examples include Skt. Siha-puras ‘Lion City’ (Singapore), Ger. Blut-wurst ‘blood sausage’.

b. Exocentric or possessive compounds (usually called bahuvrihis, see below), illustrated by Eng. redthroat, is more than the sum of its parts and refers to something outside itself: the referent is not a type of throat, but a type of bird possessing a red throat. Examples include O.Ir. Fer-gus ‘hero-strength = having a hero’s strength’, Gk. Aristó-dēmos ‘best people = having the best people’.

It is frequently said that bahuvrihis typically have o-grade of the ablauting syllable of the second compound member. Such is indeed the case in such forms as Gk. eu-pátōr (<su-patōr) ‘having a good father’, and Umbrian du-purs- (<dwi-pods) ‘having two feet’. These o-grades are likely to be survivals of old ablauting inflection rather than engendered directly by the process of compounding.

3. In the derivation of compounds special compounding rules apply.

The verbal compounds in a language observe the basic order patterns, For PIE we would expect an older OV order in compounds, as e.g. Skt. agnídh- ‘priest’ < agnifire’ + idhkindle.’

NOTE. A direct relationship between compounds and basic syntactic patterns is found only when the compounds are primary and productive. After a specific type of compound becomes established in a language, further compounds may be constructed on the basis of analogy, for example Gk. híppagroswild horse’, in contrast with the standard productive Greek compounds in which the adjectival element precedes the modified, as in agriókhoiros ‘wild swine’ (Risch 1944-1949). Here we will consider the primary and productive kinds of compounds in PIE.

4. Two large classes and other minor types are found:

A. The synthetics (noun+noun), which make up the majority of the PIE compounds:

a. Pure synthetics, i.e. noun+noun.

NOTE. While both members of the compound can be changed without a change in meaning, some rules govern these compounds (Mendoza 1998);

-  phonetic: there is a preference for a succession of high vowels – low vowels, and for labial or labialised sounds in the second term (Wackernagel 1928).

-  prosodic: the law of growing members, i.e. the tendency to place the shorter member before the longer one; that rule is followed in dvāndvā (copulative compounds of two names in the dual number, cf. Skt. Mitrā-Váruā, ‘Mitra and Varuna’, Dyvā-pr̥hivī ‘sky and earth’), but also concerning paratactic members, cf. Gk. álloi mèn rha theoí te kaì anéres híppokorustaí (Behagel 1923).

-  semantic: since Krause (1922) it is believed that the most important, near and logical notions are put in first place; i.e. compounds are subject to the principles of ‘importance’ and ‘sequence’ (Bednarczuck 1980).

b. Sinthetics in which the first element is adverbial, i.e. adverb+noun.

B. The bahuvrihis.

c. Adjective + nouns, apparently not so productive in PIE as in its dialects.

d. A small number of additive compounds.

5. The second term of a compound word may be then (Ramat 1993, Adrados–Bernabé–Mendoza 1995-1998):

i) A noun (Gk. akró-polis ‘high city, citadel’)

ii) An adjective (Gk. theo-eíkelos ‘similar to the gods’) or

iii) A noun adapted to the adjectival inflection (Gk. arguró-toxos ‘silver arc’)

NOTE. Sometimes a suffix is added (cf. Gk. en-neá-boios ‘of nine cows’), and the compound noun may have a different gender than the second term (cf. Lat. triuium ‘cross roads’, from trēs and uia).

6. The first term is a pure stem, without distinction of word class, gender or number. It may be an adverb, a numeral (Gk. trí-llistos ‘supplicated three times’, polú-llistos ‘very supplicated’) or a pronoun (cf. O.Ind. tat-purua ‘that man’), as well as a nominal-verbal stem with nominal (Gk. andra-phónos ‘who kills a man’), adjetival (Gk. akró-polis), or verbal function (Gk. arkhé-kakos ‘who begins the evil’), and also an adjective proper (Gk. polú-tropos ‘of many resources’).

7. Usually, the first term has zero-grade, cf. O.Ind. ṇr-hán, Gk. polú-tropos, Lat. aui-(caps), etc. Common exceptions are stems in -e/os, as Gk. sakés-palos ‘who shakes the shield’ (Gk. sákos ‘shield’), and some suffixes which are substituted by a lengthening in -i, cf. Gk. kudi-áneira ‘who glorifies men’ (Gk. kudrós), Av. bərəzi-čaxra- ‘of high wheels’ (Av. bərəzant-).

In thematic stems, however, the thematic -e/o appears always, as an o if noun or adjective (Gk. akró-polis), as an e if verb (Gk. arkhé-kakos).

8. The first term usually defines the second, the contrary is rare; the main compound types are:

A. Formed by verbs, cf. O.Ind. ṇr-hán, Gk. andra-phónos (Gk. andro- is newer) Lat. auceps, O.Sla. medv-ĕdĭ ‘honey-eater’, bear, and also with the second term defining the first, as Gk. arkhé-kakos.

B. Nominal determiners (first term defines the second), with first term noun (cf. Gk. mētro-pátōr ‘mother’s father’, Goth. þiudan-gardi‘kingdom’), adjective (cf. Gk. akró-polis, O.Sla. dobro-godŭ‘good time’, O.Ir. find-airgit ‘white plant’, Lat. angi-portus ‘narrow pass’), or numeral (cf. Lat. tri-uium, from uia, Gk. ámaza ‘chariot frame’, from ázōn).

C. Adjectival determiners (tatpurua-for Indian grammarians), with first term Noun (cf. Gk. theo-eíkelos, Goth. gasti-gods ‘good for the guests’), adverb (cf. O.Ind. ájñātas, Gk. ágnotos ‘unknown’, phroudos ‘who is on its way’, from pró and odós).

D. Possessive compounds (bahu-vrihi- ‘which has a lot of rice’, for Indian grammarians), as in Eng. barefoot, ‘(who goes) with bare feet’, with the first term Noun (cf. Gk. arguró-tozos, O.Sla. črŭno-vladŭ, ‘of black hair’), adjective (cf. Lat. magn-animus, ‘of great spirit’), adverb (cf. O.Ind. durmans, GK. dus-menḗs, ‘wicked’).

9. The accent could also distinguish determiners from possessives, as in O.Ind. rāja-putrás ‘a king’s son’,  from O.Ind. rajá-putras ‘who has a son as king, king’s father’, see below.


Synthetics consist of a nominal element preceding a verbal, in their unmarked forms, as in Skt. agnídh- ‘priest’. As in this compound, the relation of the nominal element to the verbal is that of target.

The particular relationship of nominal and verbal elements was determined by the lexical properties of the verb; accordingly, the primary relationship for most PIE verbs was that of target. But other nominal categories could also be used with verbs.

Kinds of relationships (Lehmann 1974):

1) The receptor relationship, as Skt. deva-héana ‘god-angering’, in m karma  devahéanam (lit. not we-do god-angering) ‘we will not do anything angering the gods’.

2) The instrument or means relationship; as Skt. ádri-jūta ‘stone-speeded’, in rátho ha vām r̥taj ádrijūta pári dyvāprthiv yāti sadyá (lit. chariot Ptc. your born-at-right-time speeded-by-stones about heaven-earth goes in-one-day) ‘Your chariot, created at the right time, speeded by stones, goes around heaven and earth in one day’.

3) The time relationship; as r̥ta-j ‘truly-born’, in the example above.

4) The source relationship, as Skt. aho-múc ‘distress-delivering’, in bháreshv índra suháva havāmahe ‘homúca sukŕ̥tam (lit. in-battles Indra well-called we-call-on freeing-from-trouble doing-well) ‘In battles we call on Indra, whom it is well to call, who frees from troubles, who does well’.

5) The place relationship, as Skt. dru-ád ‘tree-sitting’, in vér ná druác camvòr sadad dhárih (lit. bird like sitting-in-tree bowls he-has-sat fallow) ‘Like a bird sitting in a tree the fallow one has sat down in two bowls’.

6) The manner relationship; as, Skt. īśāna-kŕt ‘ruler-acting’, in ádhā yó víśvā bhúvanābhí majmánā īśānakŕt právayā abhy ávardhata (lit. here who all worlds-above with-strenght acting-like-a-ruler with-youthful-strength above he-grew) ‘Who grew beyond all worlds with his strength, acting like a ruler, having youthful strength’.

These compounds exhibit the various relationships of nominal constituents with verbal elements, as in Skt. tv-datta ‘(by-)you-given’, in vidm hí yás te adrivas tv-datta (lit. we-know Ptc. which your having-the-stones given-by-you) ‘For we know your [wealth] given by you, you of the pressing-stones’.

Synthetics attested in the Rigveda accordingly illustrate all the nominal relationships determinable from sentences. Synthetics are frequently comparable to relative constructions, as in the following sentence: gnír agāmi bhrato vtrah purucétaa (lit. to-Agni he-was-approached the-Bharatan Vrtra-killer by-many-seen) ‘Agni, the god of the Bharatas, was approached, he who killed Vṛtra, who is seen by many’.

Besides the large number of synthetics of the NV pattern, others are attested with the pattern VN. These are largely names and epithets, such as ṣṭi-gu, a name meaning ‘one who raises cattle’, and sanád-rayi ‘dispensing riches’.



The second large group of PIE compounds, Bahuvrihis, are derived in accordance with the sentence pattern expressing possession. This pattern is well known from the Latin mihi est construction (Bennett 1914; Brugmann 1911): nulli est homini perpetuom bonum ‘No man has perpetual blessings’.

NOTE. Lehmann (1974) accounts for the derivation of bahuvrihis, like Lat. magnanimus ‘great-hearted’, assuming that an equational sentence with a noun phrase as subject and a noun in the receptor category indicating possession is embedded with an equivalent noun, as in ‘great spirit is to man’ = ‘the man has great spirit’.

On deletion of the equivalent NP (homini) in the embedded sentence, a bahuvrihi compound magnanimus ‘greathearted’ is generated. This pattern of compounding ceased to be primary and productive when the dialects developed verbal patterns for expressing possession, such as Lat. habeo ‘I have’.

Bahuvrihis may be adjectival in use, or nominal, as in the vocative use of sūnari ‘having good strength’ (made up of su ‘good’ and *xner- ‘(magical) strength’) in Skt. víśvasya hí pranajvanatvé, ví yid uchási sūnari (of-all Ptc. breath life in-you Ptc. when you-shine having-good-strength) ‘For the breath and life of everything is in you, when you light up the skies, you who have good strength’. The Greek cognate may illustrate the adjectival use: phéron d’ eunora khalkón (they-bore Ptc. powerful bronze) ‘They carried on board the bronze of good strength’. The bahuvrihis are accordingly similar to synthetics in being comparable to relative clauses (Lehmann 1974).

NOTE. Although the bahuvrihis were no longer primary and productive in the later dialects, their pattern remained remarkably persistent, as we may note from the various philo- compounds in Greek, such as philósophos ‘one who holds wisdom dear’, phíloinos ‘one who likes wine’, and many more. Apart from the loss of the underlying syntactic pattern, the introduction of different accentual patterns removed the basis for bahuvrihis. As Risch pointed out, Greek eupátōr could either be a bahuvrihi ‘having a good father’ or a tatpurusha ‘a noble father’. In the period before the position of the accent was determined by the quantity of final syllables, the bahuvrihi would have had the accent on the prior syllable, like rja-putra ‘having kings as sons’, RV 2.27.7, in contrast with the tatpurusha rja-putrá ‘king’s son’, RV 10.40.3. The bahuvrihis in time, then, were far less frequent than tatpurushas, of which only a few are to be posited for Late Indo-European. An example is Gk. propátōrforefather’. If the disputed etymology of Latin propriusown’ is accepted, *pro-pətrjósfrom the forefathers’, there is evidence for assuming a PIE etymon; Wackernagel (1905) derives Sanskrit compounds like prá-padatip of foot’ from PIE. Yet the small number of such compounds in the early dialects indicates that they were formed in the late stage of PIE (Risch). Contrary to Pokorny’s reconstruction of *pro-pətrjós is the existence of adverb *proprītim (<*priH-ti-) ‘particularly’, found in Lucretius, which makes it difficult for the haplology to be sustained. A more plausible reconstruction is from verb prijājō, from prijós, dear, with an ancient meaning of ‘inalienable possesion’ in Sanskrit, or just ‘possesion’ in Latin (see Blanc 2004).

NOTE 2. Dvandvas, such as índrāviśu and a few other patterns, like the teens, were not highly productive in PIE, if they are to be assumed at all. Their lack of productiveness may reflect poorly developed coordination constructions in PIE (Lehmann 1969). Besides the expansion of tatpurushas and dvandvas in the dialects, we must note also the use of expanded root forms. Thematic forms of noun stems and derived forms of verbal roots are used, as in Skt. deva-kta ‘made by the gods’. Such extended constituents become more and more prominent and eventually are characteristic elements of compounds, as the connecting vowel -o- in Greek and in early Germanic; Gk. Apolló-dōros ‘gift of Apollo’ (an n- stem) and Goth. guma-kunds ‘of male sex’ (also an n- stem). Yet the relationships between the constituents remain unchanged by such morphological innovations. The large number of tatpurushas in the dialects reflects the prominence of embedded-modifier constructions, as the earlier synthetics and bahuvrihis reflected the embedding of sentences, often to empty noun nodes. As noted above, they accordingly have given us valuable information about PIE sentence types and their internal relationships.

10.4.3. Determiners in Nominal Phrases

Nouns are generally unaccompanied by modifiers. Demonstratives are infrequent; nouns which might be considered definite have no accompanying determinative marker unless they are to be stressed, in which case the demonstrative precedes. (Lehmann 1974).

The relationship between such demonstratives and accompanying nouns has been assumed to be appositional; it may be preferable to label the relationship a loose one, as of pronoun or noun plus noun, rather than adjective or article plus noun.

In Homer too the “article” is generally an anaphoric pronoun, differing from demonstratives by its lack of deictic meaning referring to location (Munro). Nominal phrases as found in Classical Greek or in later dialects are subsequent developments; the relationship between syntactic elements related by congruence, such as adjectives, or even by case, such as genitives, can often be taken as similar to an appositional relationship (Meillet 1937).

To illustrate nominal phrases, cf. Vedic eām marútām ‘of-them of-Maruts’. The nominal phrase which may seem to consist of a demonstrative preceding a noun, eām marútām, is divided by the end of the line; accordingly eām must be interpreted as pronominal rather than adjectival.

Virtually any line of Homer might be cited to illustrate the absence of close relationships between the members of nominal phrases; cf. Odyssey nēûs dé moi hd’ héstēken ep’ agroûnósphi pólēos, en liméni Rheíthrōi hupò Nēíōi hulenti ‘My ship is berthed yonder in the country away from the city, in a harbor called Rheithron below Neion, which is wooded’. The nouns have no determiners even when, like nēus, they are definite; and the modifiers with liméni and Neíoi seem to be loosely related epithets rather than closely linked descriptive adjectives.

The conclusions about the lack of closely related nominal phrases may be supported by the status of compounds in PIE. The compounds consisting of descriptive adjectives + noun are later; the most productive are reduced verbal rather than nominal constructions. And the bahuvrihis, which indicate a descriptive relationship between the first element and the second, support the conclusion that the relationship is relatively general; rājá-putra, for example, means ‘having sons who are kings’ rather than ‘having royal sons’; gó-vapus means ‘having a shape like a cow’, said of rainclouds, for which the epithet denotes the fructifying quality rather than the physical shape. (Lehmann 1974).

Accordingly, closely related nominal expressions are to be assumed only for the dialects, not for PIE. Definiteness was not indicated for nouns. The primary relationship between nominal elements, whether nouns or adjectives, was appositional.

10.4.4. Apposition

1. Apposition is traditionally “when paratactically joined forms are grammatically, but not in meaning, equivalent” (Lehmann 1974).

NOTE. Because of the relationship between nouns and modifiers, and also because subjects of verbs were only explicit expressions for the subjective elements in verb forms, Meillet (1937) considered apposition a basic characteristic of Indo-European syntax. Subjects were included only when a specific meaning was to be expressed.

2. A distinction is made between appositional and attributive (Delbrück); an appositional relationship between two or more words is not indicated by any formal expression, whereas an attributive relationship generally is (Lehmann 1974).

A. Thus the relationships in the following line of the Odyssey are attributive: arnúmenos hn te psukhn kaì nóston hetaírōn lit. ‘striving-for his Ptc. life and return of-companions’. The relationship between hn and psukhn is indicated by the concordance in endings; that between nóston and hetaírōn by the genitive.

B. On the other hand the relationship between the two vocatives in the following line is appositional, because there is no mark indicating the relationship: tȏn hamóthen ge, theá, thúgater Diós, eipè kaì hēmȋn ‘Tell us of these things, beginning at any point you like, goddess, daughter of Zeus’. Both vocatives can be taken independently, as can any appositional elements.

3. Asyndetic constructions which are not appositive are frequently attested, as Skt. té vo h mánase santu yajñ ‘These sacrifices should be in accordance with your heart, your mind’. Coordinate as well as appositive constructions could thus be without a specific coordinating marker.

4. Comparable to appositional constructions are titles, for, like appositions, the two or more nouns involved refer to one person.

NOTE. In OV languages titles are postposed in contrast with the preposing in VO languages; compare Japanese Tanaka-san with Mr. Middlefield. The title ‘king’ with Varuna and similarly in the Odyssey, Poseidáōni ánakti, when ánaks is used as a title. But, as Lehmann himself admits, even in the early texts, titles often precede names, in keeping with the change toward a VO structure (Lehmann 1974).

5. Appositions normally follow, when nouns and noun groups are contiguous, as in the frequent descriptive epithets of Homer: Tòn d’ ēmeíbet’ épeita theá, glaukôpis Ath, ‘Him then answered the goddess, owl-eyed Athene’.

To indicate a marked relationship, however, they may precede (Schwyzer 1950). But the early PIE position is clear from the cognates: Skt. dyaus pitā, Gk. Zeû páter, Lat. Jūpiter.

10.5. Modified forms of PIE Simple Sentences

10.5.1. Coordination

1. While coordination is prominent in the earliest texts, it is generally implicit.

The oldest surviving texts consist largely of paratactic sentences, often with no connecting particles; cf. Lat. ueni, uidi, uici ‘I came, I saw, I won’; Gk. iomen, hōs ekéleues, anà drumá …/ heúromen en bssēisi tetugména dmata kalá lit. ‘we-went as you-commanded, through the-jungle…/we-found in the-valley a-polished palace beautiful’; Hitt. adueni akueni nu URUattusa iyannae ‘we-eat, we-drink Ptc. from-Hatusa I-go’.

2. New sentences may be introduced with particles, or relationships may be indicated with pronominal elements; but these are fewer than in subsequent texts.

Similar patterns of paratactic sentences are found in Hittite, with no overt marker of coordination or of subordination. According to Friedrich (1960) “purpose and result” clauses are not found in Hittite; coordinate sentences are simply arranged side by side with the particle nu, as in the Hittite Laws. Conditional relationships too are found in Hittite with no indication of subordination, v.i. §10.5.3.

The arrangement of sentences in sequence is a typical pattern of PIE syntax, whether for hypotactic or for paratactic relationships.

3. Expressions for coordination were used largely for elements within clauses and sentences. When used to link sentences, conjunctions were often accompanied by initial particles indicating the beginning of a new clause and also indicating a variety of possible relationships with neighbouring clauses (Lehmann 1974).

NOTE. Sentence-connecting particles are, however, infrequent in Vedic and relatively infrequent in the earliest Hittite texts; Lehmann (1974) concludes that formal markers of sentence coordination were not mandatory in PIE.


The normal coordinating copulative particle in most of the dialects is a reflex of PIE -qe.

NOTE. Hittite -a, -i̯a is used similarly, as in attaš annaš a ‘father and mother’ (J. Friedrich 1960). This is probably related to Toch. B yo.

This is postposed to the second of two conjoined elements; as, Ved. ágna índras ca ‘o Agni and Indra’; or to both, as Skt. devś ca ásurās ca, ‘Gods and Asuras’, Gk. patr andrõn te theõn te ‘father of men and gods’ (Beekes 1995).

Introducing another sentence, as Gothic fram-uh þamma sokida Peilatus fraletan inaAnd at this Pilate sought to release him’.

There is an tendency toward a polysyndetic use (Mendoza 1998); cf. Gk. aieì gár toi épis te phílē pólemoi te mákhai te (lit. always then for-you the-discord -and dear the-wars -and the-battles -and) ‘Discord, wars and battles are always dear to you’.

NOTE. With the change in coordinating constructions, new particles were introduced; some of these, for example, Lat. et, Goth. jah, O.E. and, have a generally accepted etymology; others, like Gk. kaí, are obscure in etymology. Syntactically the shift in the construction rather than the source of the particles is of primary interest, though, as noted above, the introduction of new markers for the new VO patterns provides welcome lexical evidence of a shift. The syntactic shift also brought with it patterns of coordination reduction (Ersparung) which have been well described for some dialects (Behaghel). Such constructions are notable especially in SVO languages, in which sequences with equivalent verbs (S, V, O, Conj., S2, V1, O2) delete the second occurrence of the verb , as M.H.G. daz einer einez will und ein ander ein anderz, ‘that one one-thing wants and another another’. Lehmann (1974)


The disjunctive particle PIE -w is also postposed to the second element; Lat. silua alta Iouis lūcusue Diānae ‘the high forest of Jupiter or the grove of Diana’; or to both, as Skt. nákta vā dívā vā, ‘(either) during the night or during the day’, or Gk. ḕ theòs ēè gun (with change to prepositional order) ‘or goddess or woman’.

NOTE. In Hittite, however, the postposed particles -ku ... -ku (<-qe…-qe) ‘or’, were used with alternative function, always repeated and only as nexus between sentences; for nouns there was the particle našma, which stood between nouns rather than after the last. This pattern of conjunction placement came to be increasingly frequent in the dialects; it indicates that the conjunction patterns of VO structure have come to be typical already by PIH.


In all older IE languages it was possible to express this function of parataxis by simple repetition of negation, and that was the only possibility in Vedic.

However, a disjunctive compound conjunction neqe is found in Indo-Iranian, Anatolian, Italic, Celtic and Germanic. It was not a frozen lexical remain, since the older IE languages analyse the compound as ne+qe.

NOTE. A compound with the prohibitive negation mqe is also found in Indo-Iranian and Greek, and with the alternative newe in Indo-Iranian, Anatolian, Italic and Celtic.

The compound conjunction can appear introducing only the second member of both negatives, or it can be repeated also in the first member to emphasise the parallelism of this type of construction (Mendoza 1998).

Both systems, as well as the simple negation, are attested in the oldest texts.  


No common adversative conjunction seems to be reconstructible for PIE, although dialectally some conjunctions did succeed with this function, probably in a LIE or post-LIE period; as, at(i), ōd, etc. (v.s. §8.5).

NOTE. That has been related by Bednarczuk (1980) with the postpositive character of paratactic conjunctions of the common language, what configures them as polysyndetic and capable of joining words and sentences, while the adversative function – exclusive of the sentence parataxis – demands prepositive conjunctions and necessarily monosyndetic.

In the older IE languages, the adversative function is expressed by the same postposed copulative conjunction -qe, being its adversative use defined by the context, by the semantic relationship between the sentences (Gonda 1954).

10.5.2. Complementation

1. Compound sentences may result from the embedding of nominal modifiers.

NOTE. In VO languages embedded nominal modifiers follow nouns, whereas in OV languages they precede nouns. This observation has led to an understanding of the Hittite and the reconstructed PIE relative constructions. If we follow the standard assumption that in relative constructions a second sentence containing a noun phrase equivalent to a noun phrase in the matrix sentence is embedded in that matrix sentence, we may expect that either sentence may be modified. A sentence may also be embedded with a dummy noun; the verb forms of such embedded sentences are commonly expressed with nominal forms of the verb, variously called infinitives, supines, or participles. In OV languages these, as well as relative constructions, precede the verb of the matrix sentence (Lehmann 1974).

2. An example with participles in the IE languages is Skt. vásāna in the last lines of the following strophic hymn: rúśad vásāna sudśīkarūpa lit. ‘brightly dressing-himself beautifully-hued’.

It may also have “a final or consequential sense”, as in the following strophic hymn: tvám indra sravitav apás ka ‘You, O Indra, make the waters to flow’. Also in the poetic texts such infinitives may follow the main verb, as in ábodhi hótā yajáthāya devn (lit. he-woke-up priest for-sacrificing gods) ‘The priest has awakened to sacrifice to the gods’ (Lehmann 1974).

NOTE. The postposed order may result from stylistic or poetic rearrangement; yet it is also a reflection of the shift to VO order, a shift which is reflected in the normal position for infinitives in the other IE dialects. In the Brahmanas still, infinitives normally stand directly before the verb, except in interrogative and negative sentences (Delbrück).  On the basis of the Brahmanic order we may assume that in PIE nonfinite verbs used as complements to principal verbs preceded them in the sentence. Hittite provides examples of preposed complementary participles and infinitives to support this assumption (J. Friedrich). Participles were used particularly with har(k)- ‘have’ and eš-be’, as in uerii̯an ešta ‘was mentioned’; the pattern is used to indicate state.

Infinitives and Participles

According to Fritz (in Meier-Brügger 2003) in Proto-Indo-European, the existence of participles may safely be asserted. Additionally, infinitive constructions with final dative, accusative of direction, and the locative of destination are presumed to have existed. While infinitives are defined by syntax, the very presence of participles reveals that there were participial constructions in Proto-Indo-European. According to J.L. García Ramón (1997), “[the assertion] of a true Proto-Indo-European infinitive ending in *-sén(i) is justified,” which could be traced to a locative form, which, according to K. Stüber (2000) appears in the case of s-stem abstract nouns with the locative forms ending in *en.

Infinitives could indicate result, with or without an object (J. Friedrich 1960): 1-aš 1-an kunanna šanhanzi (lit. oneoneto-killnothe-tries) ‘One should not try to kill another’.

Infinitives could be used to express purpose, as in the following example, which pairs an infinitive with a noun (J. Friedrich): tuk-ma uttar ŠÀ-ta šii̯anna išhiull-a ešdu (lit. to-you-however this word in-heart for-laying instruction-and it-should-be) ‘But for you this word should be for taking to heart and for instruction’.

The infinitive could be loosely related to its object, as in examples cited by Friedrich, such as apāš-ma-mu harkanna šan(a)hta (lit. “he-however-me for-deteriorating he-sought) ‘But he sought to destroy me’.

The complementary infinitive indicates the purpose of the action; as Friedrich points out, it is attached to the verb šanhta plus its object mu in a construction quite different from that in subsequent dialects.

These uses are parallelled by uses in Vedic, as may be noted in the work of Macdonell (1916), from which some examples are taken in Lehmann (1974). On the basis of such examples in Vedic and in Hittite, he assumes that infinitive constructions were used to indicate a variety of complements in PIE.

Hittite and Sanskrit also provide examples of participles functioning appositionally or as adjectives indicating state (J. Friedrich 1960): ammuk-u̯ar-an akkantan IQ.BI (lit. to-me-Ptc.-indicating-quotation-him dying he-described) ‘He told me that one had died’.

This pattern had been noted by Delbrück (1900) for the Rigveda, with various examples, as śiśīhí śiśayá tvā śṛṇomi, ‘Strengthen me; I hear that you are strong’. The adjective śiśayá ‘strengthening’ is an adjective derived from the same root as śiśīhí. Delbrück also noted that such “appositives” are indicated in Greek by means of clauses. Greek represents for Lehmann accordingly a further stage in the development of the IE languages to a VO order. Yet Greek still maintained preposed participles having the same subject as does the principal verb, as in: tḕn mèn idṑn gthēse, lit. ‘it Ptc. seeing he-rejoiced’ (Lehmann 1974).

This pattern permits the use of two verbs with only one indicating mood and person; the nonfinite verb takes these categories from the finite.

 Participles were thus used in the older period for a great variety of relationships, though also without indicating some of the verbal categories.

Dependent clauses are more flexible in indicating such relationships, and more precise, especially when complementary participles and infinitives follow the principal verb.


10.5.3. Subordinate Clauses

1. As with coordination, subordination could be made without grammatical mark in the older IE languages, hence the context alone decided if it was a subordinate; as in Eng. the man he called paid the boy (Lehmann 1974).

So e.g. Hom. Gk. all’ áge nûn epímeinon, aria teúkhea duō ‘then now wait (until) I dress the weapons of war’; Lat. fac noscam ‘make me know’ (Delbrück 1900).

2. These sentences, with a change in person, mood, or simply eliminating the pause between both sentences, are usually considered the origin of some types of subordinates.

Especially common is this archaic type of paratactic construction in substantive subordinates, i.e. those sentences which take the role of a verbal actor (subject or object), or of a constituent of the nominal sentence. Such archaic forms are found in substantive subordinates expressing thoughts, words, desires or fears, in which these are quoted literally, without change in person, i.e. in the direct syle, proper of a stage of the language previous to the introduction of the indirect style (Rubio 1976).

3. The most extended substantive sentence in the older languages is in fact not the one introduced by conjunctions – which is considered a late development –, but those whose subordination is introduced by non-finite forms of the verb, like participles and infinitives.

Almost all languages have infinitives and participles functioning as subject or object of the sentence, especially with verbs of will or in nominal sentences. That is found in Hittite, Greek, Latin, Germanic, Slavic, and especially frequent are participles in -nt (the oldest ones), clearly differentiated from the main clause with the common subordinate+main order.

NOTE. The order subordinate+main seems to have been obligatory in PIE. This has been linked with the lack of subordinates of purpose and result, because in these constructions iconicity demands a postposed order. Precisely conditionals – apparently the first subordinates introduced by conjunctions, see below – have a structure which is necessarily correlative, with a systematic preposed order to the subordinate.

In the attested IE languages that relative order subordinate+main was eventually inverted, which has been linked with a typological change of a language OV into a VO one. At the same time, there was a tendency to place the relative pronoun immediately after the noun to which it refers, all of which is typically found in most IE languages of a late period.

Examples of such subordinate clauses without mark include (Mendoza 1998): Hitt. MU.KAM-za-wa-ta šer tepawešanza nu-wa BE-LI-NI INA URUayaša lē paiši ‘the year having been shortened, (Ptc.) Lord, do not go to Hayasa’; Gk. kaì prín per thumō̑i memas Tróessi mákhesthai dè tóte min trìs tósson hélen ménos lit. ‘and before even in-his-soul willing-to against-the-Trojans fight Ptc. then to-him thrice bigger took-him impetus’.

NOTE. According to Mendoza (1998), such examples are usually analysed as absolute constructions (see above §10.1.2). In Hittite, where there are no absolutes, they are interpreted as pure nominal sentences, without copula.

But, if the relation between both clauses is inverted, and with it their order, the result is two juxtaposed sentences, the second one (with participle) introduced by nu too, but with the personal verb, with an explicit copula: Hitt. man I-NA URUayaša paun-pat nu-za MU.KAM-za-wa-ta šer tepawešanza ešta (lit. ptc. to Hayasa I-would-have-gone ptc. the-year in-which-it-has-been-shortened is) ‘I would have gone to Hayasa, but the year got shorter’.

The conclusion is that from these sentences, with the integration of such substantive subordinate constructions into the main clause, come the absolute participle constructions attested in all IE branches but for Anatolian (Holland 1986).

4. A common resource in the older IE languages was to ‘transform’ finite verbs into nominal sentences, due to the syntactic flexibility of participles. As, for Gk. hoppóteros dé ke niksēi … gunaĩká te oíkad’ agésthōwho wins … he carries home the woman’, Gk. tõi dé ke niksanti phílē keklsēi ákoitisby the one having won you will be called dear wife’ (Ramat 1993).

NOTE. So e.g. the performative assertion, a linguistic act of guarantee and compromise, expressed by the syntactic secondary rule of demonstrative in Ved. ayám te asmi ‘with-that yours I-am’, is transformed into a participle clause in the complex m mm imam táva sántam … ni gārit ‘that he bolts me not’, lit. ‘being here yours’. The participle reinforces a performative assertion ‘With that I  promise you, Atri. Do not let Svarbhānu destroy me’.  This is the case of the finite verb of existence in another linguistic act, of confessional formula; as, Hitt. ēšziy=at iyawen=at ‘It is. We did it’, which is transformed into ašān=at iyanun=at , lit. ‘This (is) absent. I did it’. From this syntactic use of Hitt. ašant- we can glimpse Lat. sōns, sontis ‘guilty’, the old participle present of verb ‘be’, esse.  Ramat (1993).


Of all subordinate clauses introduced by conjunctions, only conditionals seem not to be related to the relative clauses. In fact, they are the only conjunctional subordinate developed in Hittite, well attested since the older texts (introduced by takku), and whose conjunction is not derived from the relative (Mendoza 1998).

According to Clackson (2007), a particular marker appears to be used to introduce conditional clauses in at least three different early branches of IE: -qe, the connective enclitic (although it also has other functions), both at the word level and, less generally, clause level. The Sanskrit derivative of -qe, ca, is used in thirty-one passages in the Rig-Veda, the corpus of early Sanskrit hymns, to introduce subordinate conditional or temporal clauses; in all cases the clause with the clitic ca stands before the main clause.

Wackernagel (1942) had noticed similar, marginal, uses of the cognates of Sanskrit ca to introduce subordinate, and in particular conditional, clauses in Gothic, Latin and Greek. This support for a PIE use of -qe to introduce conditional clauses then appeared to be confirmed by Hittite, since the element -ku of the conditional particle takku ‘if’, can be derived from -qe.


Indo-Europeanists have long recognised the relationship between the subordinating particles and the stem from which relative pronouns were derived in Indo-Iranian and Greek.

Thus Delbrück has pointed out in detail how the neuter accusative form of PIE jo- was the basis of the conjunction jod in its various meanings: (1) Temporal, (2) Temporal-Causal, (3) Temporal-Conditional, (4) Purpose. He also recognised the source of conjunctional use in sentences like Skt. yáj jyathās tád áhar asya kme śó pīyū́am apibo giriṣṭhm, ‘On the day you were born you drank the mountain milk out of desire for the plant’ (Lehmann 1974).

Thus, subordinated clauses that are introduced by relative pronouns can perform the function of subject, object, adverbial phrase, appositional phrase, and attribute.

a) In the older IE languages, the relative clause often precedes the main clause (and with it, the antecedent). The relative pronoun or adverb is often paired with a pronominal or adverbial antecedent, yielding what are called correlative structures of the type ‘(the one) who …, he …’, or ‘in the way which …, in that way…’ (Fortson 2004).

b) The earliest type of subordinate jo-, qi-/qo- clauses must have been the preposed relative constructions; Vedic no dvéi, ádharah sás padīawho us hates, down he will-fall’. But elements could be left-detached for topicalisation (see below §10.6.3): púro yád asya sampinak (lit. rocks when of-him you-destroyed) ‘when you destroyed his rocks’.

NOTE. This conclusion from Vedic receives striking support from Hittite, for in it we find the same syntactic relationship between relative clauses and other subordinate clauses as is found in Vedic, Greek, and other early dialects. But the marker for both types of clauses differs. In Hittite it is based on IE qid rather than jod; thus, Hittite too uses the relative particle for indicating subordination. The remarkable parallelism between the syntactic constructions, though they have different surface markers, must be ascribed to typological reasons; we assume that Hittite as well as Indo-Aryan and Greek was developing a lexical marker to indicate subordination. As does yad in Vedic, Hitt. kuit signals a “loose” relationship between clauses which must be appropriately interpreted. Lehmann (1974).

As J. Friedrich has stated (1960), kuit never stands initially in its clause. Sentences in which it is used are then scarcely more specifically interconnected than are conjoined sentences with no specific relating word, as in examples cited by Friedrich (ibid.): nu taškupāi nu URU-aš dapii̯anzi išdammašzi (lit. Ptc. you-shout Ptc. city whole it-hears) ‘Now cry out [so that] the whole city hears’. Like this example, both clauses in a kuit construction generally are introduced with nu (J. Friedrich 1960). We may assume that kuit became a subordinating particle when such connections were omitted, as in Friedrich’s example. These examples illustrate that both yád and kuit introduce causal clauses, though they do not contain indications of the origin of this use.


It is therefore generally believed that subordinates originated in relative sentences, as Vedic, Old Irish, Avestan and Old Persian illustrate. Proverbs and maxims are a particularly conservative field in all languages, and even etymologically there are two series which appear especially often in correlatives; namely, qo-...to-, and jo-...to-.

NOTE. For IE qo-..to-, cf. Lat. cum...tum, qualis...talis, quam...tam, or Lith. kàs...tàs, kòks...tàs, kaîp...taîp, kíek...tíek, etc., and for jo-...to-, Ved. yás... tád, yáthā...táthā, yvat...tvat, Gk. oios...toios, ósos...tósos, O.Pers. haya (a compound from so+jo, with the same inverse compound as Lat. tamquam, from two correlatives), etc.

For Haudry this correlative structure is the base for subordination in all Indo-European languages. Proto-Indo-European would therefore show an intermediate syntax between parataxis and hypotaxis, as the correlative structure is between a ‘loose’ syntax and a ‘locked’ one.

Examples of equivalent old correlative sentences include the following (Watkins 1976, Ramat 1993): Hitt. nu tarzi kuiš nu apāš KA.TAB.ANŠE ēpzi (lit. Ptc. wins who Ptc. he reins takes) ‘Who wins, (he) takes the reins [=takes charge]’; Ved. sa yo na ujjeyati sa pratama somasya pāsyati lit. ‘he who Ptc. shall-win, he the-first the-soma will-drink’; Ved. sa yo na ujjeyati tasya idam bhaviyatihe who Ptc. shall-win, of-him this will-be’; Gk. hós nun orkhéstōn … atalótata paízei tóde k[]nwho now of the dancers more sporting plays, of him (is) this […]’; Gk. hoppóteros dé ke niksēi … gunaĩká te oíkad’ agésthōwho wins … he carries home the woman’.

Restrictive and Explicative Relative Clauses

Greek, Indo-Iranian, Phrygian, Slavic and Celtic have inherited a stem jo- but Anatolian, Latin, Sabellian and Tocharian derive their relative pronouns from a stem qo-/qi-. This distribution cuts across other isoglosses separating the IE languages and does not seem to reflect a dialectal difference of the parent language.

Common examples of relative clauses are (Fortson 2004): yéna im viśvā cyavanā kr̥tni ... sá janāsa índra ‘(The one) by whom all these things have been made to shake ... that, people, (is) Indra’; Gk. hós ke theoĩs epipeíthētai, mála t’ ékluon autoĩWhoever obeys the gods, they listen to him as well’.

NOTE. Fritz (in Meier-Brügger 2003) sums up the uses of the qi-/qo- and jo- relative pronouns, according to Ch. Lehmann. Clackson’s (2007) description is very clear: “A crucial element of the recent work has been the difference between restrictive or defining relatives and non-restrictive (also called descriptive or appositional) relatives. Restrictive relatives delimit the head of the relative clause, but non-restrictive relatives merely add extra information about their head. Compare the following sentences:

[a] The tea that I drank was cold.

[b] The tea, which I eventually drank, was cold.

In [a] the relative defines and restricts the referent, whereas in [b] the relative gives incidental information, and is in effect a separate assertion from that of the main clause. In English, if the relative pronoun is omitted, only a restrictive interpretation is possible. Some English speakers prefer, or are taught, to use that as a relative pronoun in restrictive clauses.”

The differentiation of relative clauses introduced by qo-/qi- and jo- is summarised by Clackson (2007) according to the finds in the older IE languages:

·  qo-/qi- shows a common use for attributive-restrictive relative clauses in Hittite and Latin, and their most common order is relative-matrix; as, Lat. pecuniam quis nancitor habeto ‘fortune, who acquire it own it’, Goth. ni weistu hwaz ih sagen thir, ‘you don’t know what I say to you’.

·  jo- is most commonly used in appositive-explicative relative clauses in Vedic Sanskrit and Homeric Greek, with a matrix-relative order of the sentence; as, Ved. sóma ys te mayobhúva ūtáyah sánti dāśúe tbhir no ‘vit bhava (lit. Soma which from-you benefitious aids are for-the-one-who-worships-you, with-them of-us helper be-you) ‘Soma, with your aids, which are benefitious to those who worship you, help us’.

NOTE. Clackson (2007): “Should we then conclude that PIE had two separate relative pronouns, and different clause structures for restrictive and non-restrictive relatives? It is often a fault of Indo-Europeanists to over-reconstruct, and to explain every development of the daughter languages through reconstruction of a richer system in the parent language. (…) However, the reconstruction of two relative pronouns for PIE does fit the attested facts better than any of the other theories on offer.”

Very characteristically, if the antecedent is a noun rather than a pronoun, it is placed within the relative clause and in the same case as the relative, sometimes repeated in the main clause. Thus instead of saying The gods who gave us riches can take them away, speakers of these languages would have said literally, Which gods gave us riches, they/those gods can take them away (Fortson 2004): Hitt. nu=kan kāš IM-aš kuēz wappuwaz danza nu zik wappuaš DMA tuēl ŠU-TIKA dā (lit.) ‘from which riverbank this clay (has been) taken, o genius of (that) riverbank, take (it) in your hand’, i.e. ‘0 genius of the riverbank from which this clay has been taken ...’; Old Latin quem agrum eōs uēndere herēdemque sequī licet, is ager uectīgal nei siet ‘the field which (lit., which field) they are allowed to sell and pass to an heir, that field may not be taxable’.

NOTE. As can be seen from some of the examples so far quoted, the relative pronoun did not need to be the first member of its clause. In several of the ancient IE languages, the relative could be preceded at least by a topicalised element, just like the subordinating conjunctions.


Lehmann (1974) assumes that the use of Skt. yád, Hitt. kuit, and other relative particles to express a causal relationship arose from subordination of clauses introduced by them to an ablative; cf. Skt.  ácittī yát táva dhármā yuyopimá lit. ‘unknowingthat, becauseyourlaw, orderwe-have-disturbed’, m nas tásmād énaso deva rīria (lit. not us because-of-that because-of-sin O-god you-harm) ‘Do not harm us, god, because of that sin [that] because unknowingly we have disturbed your law’.

As such relationships with ablatives expressing cause were not specific, more precise particles or conjunctions came to be used. In Sanskrit the ablatival yasmāt specifies the meaning ‘because’.

Further, yad and yátra specify the meaning ‘when’. In Hittite, mān came to be used for temporal relationships, possibly after combined use with kuit; kuitman expressed a temporal relationship even in Late Hittite, corresponding to ‘while, until’, though mahhan has replaced mān (J. Friedrich 1960 gives further details). The conjunction mān itself specifies the meanings ‘if’ and ‘although’ in standard Hittite. In both Hittite and Vedic then, the “loose” relative-construction relationship between subordinate clauses and principal clauses is gradually replaced by special conjunctions for the various types of hypotactic relationship: causal, temporal, conditional, concessive.

Just as the causal relationship developed from an ablative modified by a relative construction, so the temporal and conditional relationship developed from a clause modifying an underlying time node.

10.6. Syntactic Categories

10.6.1. Particles as Syntactic Means of Expression

Noninflected words of various functions were used in indicating relationships between other words in the sentence or between sentences.

1. Some were used for modifying nouns, often indicating the relationships of nouns to verbs. Although these were generally placed after nouns and accordingly were postpositions, they have often been called prepositions by reason of their function rather than their position with regard to nouns (Delbrück).

2. Others were used for modifying verbs, often specifying more precisely the meanings of verbs; these then may be called preverbs.

3. Others, commonly referred to as sentence connectives, were used primarily to indicate the relationships between clauses or sentences (Watkins 1964; Lehmann 1969).

Prepositions and Postpositions

Prepositions and postpositions were simply independent adverbs in PIE (as in Anatolian, Indo-Iranian and the oldest Greek), and they could appear before or after their objects, although the oldest pattern found is agreed to be postposed. Anatolian and Vedic have almost exclusively postpositions, not prepositions; as, Hitt. šuḫḫi šēron the roof’, Ved. jánām̐ ánuamong men’; and also remains in Gk. toútōn périabout these things’, and Lat. mēcumwith me’ (Fortson 2004).

Postpositions in the various dialects are found with specific cases, in accordance with their meanings.

Yet in the Old Hittite texts, the genitive rather than such a specific case is prominent with postpositions derived from nouns, such as piran ‘(in) front’ (Neu 1970): kuiš LUGAL-ua-aš piran ēšzi (who king’s front he-sits) ‘whoever sits before the king’ (Lehmann 1974).


Rather than having the close relationships to nouns illustrated above, the same adverbs could instead be associated primarily with verbs, often the same particles which were used as postpositions.

Examples include (Fortson 2004) Hitt. š=aš šarā URU-ya pait ‘and he went up to the city’; Ved. abhí yó mahin díva mitró babhū́va sapráthā ‘Mitra the renowned who is superior to heaven by his greatness’; O.Av. frō mā sāstū vahištā let him teach me the best things’; Gk. edētúos eks éron héntothey put aside desire for food”; O.Lat. ob uōs sacrō ‘I entreat you’ (would be uōs obsecrō in Classical Latin); O.Ir. ad- cruth caín cichither ‘fair form will be seen’.

NOTE. German and Dutch are well known for having many separable affixes. In the sentence Ger. Ich komme gut zu Hause an the prefix an in the verb ankommen is detached. However, in the participle, as in Er ist angekommen ‘He has arrived’, it is not separated. In Dutch, compare Hij is aangekomen ‘He has arrived’, but Ik kom morgen aan  ‘I shall arrive tomorrow’.

 English has many phrasal or compound verb forms that act in this way. For example, the adverb (or adverbial particle) up in the phrasal verb to screw up can appear after the subject (“things”) in the sentence: He is always screwing things up.

Non-personal forms, i.e. nouns and adjectives, form a compound (karmadharaya) with the preposition; as O.Ind. prasāda ‘favour’, Lat subsidium, praesidium, O.Ind. apaciti, Gk. apotisis ‘reprisal’, etc.

Preverbs might occupy various positions:

1. If unmarked, they are placed before the verb, as in the examples above.

2. If marked, they are placed initially in clauses (Watkins 1964); as, Av. pairi uši vāraiiaϑβəmcover (their) ears’, where the preverb (pairi, literally ‘around’) has been fronted to the beginning of the clause for prominence or emphasis.

NOTE. In the course of time the preverbs in unmarked position came to be combined with their verbs, though the identity of each element is long apparent in many of the dialects. Thus, in Modern German the primary accent is still maintained on some verbal roots, and in contrast with cognate nouns the prefix carries weak stress: ertéilen ‘distribute’, Úrteil ‘judgment’. The steps toward the combination of preverb and verbal root have been described for the dialects, for example, Greek, in which uncombined forms as well as combined forms are attested during the period of our texts.

NOTE. In the attested IE dialects:

- Preverbs which remained uncombined came to be treated as adverbs.

- Combinations of preverbs plus verbs, on the other hand, eventually came to function like unitary elements.

The two different positions of preverbs in early texts led eventually to different word classes.

Sentence Particles

Particles were also used to relate sentences and clauses (J. Friedrich 1959):






























‘If anyone breaks the hand or foot of a freeman, then he must give him twenty shekels of silver.’

Particles like the initial word in this example indicate the kind of clause that will follow and have long been well described. The function of particles like nu is not, however, equally clear.

NOTE. Dillon and Götze related nu and the use of sentence connectives to similar particles in Old Irish (Dillon 1947). Such particles introduce many sentences in Old Irish and have led to compound verb forms in this VSO language. Delbrück had also noted their presence in Vedic (1888).

Since introductory šu and ta were more frequent than was nu in the older Hittite texts, scholars assumed that sentences in IE were regularly introduced by these sentence connectives. And Sturtevant proposed, as etymology for the anaphoric pronoun, combinations of so- and to- with enclitic pronouns, as in the well-known Hittite sequence ta-at, cf. IE tod, and so on (see Otten and Souček 1969 for the use of such particles in one text).

It is clear that sentence connectives were used in Hittite to indicate continued treatment of a given topic (Raman 1973). It is also found with Hittite relative constructions, a function which may also be ascribed to Vedic and tád.

Compare this syntactic use of particles , , , in Ved. sá hovāc Gargya lit. ‘Ptc. Ptc.-said Gargyas’tád u hovācāsuri lit. ‘Ptc. Ptc. Ptc.-said-Asuri’.

NOTE. For Lehmann (1974), since this use may be accounted for through post-PIE influences, sentence connectives may have had a minor role in PIE.

Other particles, like Hitt. takku ‘if’, had their counterparts in PIE, in this case in -qe. This is also true for emphatic particles like Skt. íd; they were used after nouns as well as imperatives.

10.6.2. Marked Order in Sentences

1. Elements in sentences can be emphasised, by marking; the chief device for such emphasis is initial position, i.e. elements are moved leftward in a process called fronting.

2. In unmarked position the preverb directly precedes the verb. Changes in normal order thus provide one of the devices for conveying emphasis.

Other devices have to do with selection, notably particles which are postposed after a marked element.

3. Emphasis can also be indicated by lexical selection.

4. Presumably other modifications might also be made, as in intonation.

The various syntactic devices accordingly provided means to introduce marking in sentences.

10.6.3. Topicalisation with Reference to Emphasis

1. Like emphasis, topicalisation is carried out by patterns of arrangement, but the arrangement is applied to coequal elements rather than elements which are moved from their normal order.

2. Topicalisation by arrangement is well known in the study of the early IE languages, as in the initial lines of the Homeric poems. The Iliad begins with the noun mȇnin ‘wrath’, the Odyssey with the noun ándra ‘man’, opening both poems: mȇnin áeide ‘Sing of the wrath’ and ándra moi énnepe ‘Tell me of the man’. The very arrangement of moi and other enclitics occupying second position in the sentence, in accordance with Wackernagel’s law, indicates the use of initial placement among nominal elements for topicalisation (Lehmann 1974).

Examples (Fortson 2004) include Hitt. alziššai=wa=tta DINGIRMEŠ-aš attaš DKumarbiš ‘Kumarbi, the father of the gods, is calling you’; O.Av. sraōtū sāsnā̊ fšə̄ŋ́hiiō suiie taštō ‘Let the bondsman (?), fashioned for benefit, hear the teachings’, Gk. ménei tò theĩon doulíāi per en phrení ‘The divine (power), even when in bondage, stays in the mind’, Lat. fuimus Trōes, fuit Īlium ‘We were (but no longer are) Trojans, Troy was (but no longer is)’.

Fortson (2004): “Certain verbs, especially existential verbs  (e.g., ‘there is’) but also verbs of speaking and imperatives, preferentially occur clause-initially across all the IE languages: Skt. āsīd rājā nalo nāmathere was a king named Nala’ (Mahābhārata 3.53.1), Lat. est in cōnspectū Tenedos nōtissima fāmā īnsula ‘within sight there is a most famous island, Tenedos’ (Vergil, Aeneid 2.21-22), dialectal Old Russian estĭ gradŭ mežu nobomŭ i zemlejuthere is a city between heaven and earth’ (Novgorod birch bark fragment 10.1).”

NOTE. The fronted element was characteristic of certain situations of the discourse, like the beginning of a text, the kataphora (repetition of a cohesive device at the end of a sentence, like a pronoun catching up an antecedent) and imperatives (Ramat 1993); compare e.g. the beginning of a typical tale ‘there was a king’, in Skt. āsīd rājā, Gk. ske tis … (w)annássōn, Lith. bùvo karãlius, O.Ir. boí, Russ. žyl byl korol’, etc.

As these passages and many others that might be cited illustrate, the basic sentence patterns could be rearranged by stylistic rules, both for emphasis and for topicalisation. In this way the relatively strict arrangement of simple sentences could be modified to bring about variety and flexibility.

3. Clause-initial position is a place of prominence for any constituent, not only for verbs; as Hitt. irma=šmaš=kan dāḫḫun “sickness I have taken away from you”. Topicalisation was probably a syntactic process in PIE (Fortson 2004).

4. Interrogatives, as already stated, move forward to the so-called complementiser position, which can also be occupied by other elements, like relative pronouns or subordinating conjunctions. The complementiser position precedes the rest of the clausal positions proper, but this position is preceded by the topicalisation position; if the latter is filled by a topicalised element, the complementiser is no longer clause-initial (Fortson 2004).

Some examples are Hitt. ammuqq=a kuit arkun ‘And also (that) which I had”, Ved. jātám yád enam apáso ádhārayan ‘when the craftsmen held him, just born’, O.Av. naēnaēstārō yaϑənā vohunąm mahī ‘since we are non-scorners of good things’, fēstō diē sī quid prodēgeris ‘if you splurge a bit on a holiday’.

Topicalisation usually consists of one constituent or subconstituent, but it can be more complex: Ved. áher yātram kám apaśya indrawhich avenger of the snake you saw, o Indra?’ ádevena mánasā riayátiwho does wrong with impious intention’.

 Cf. also from the Archaic Latin of Plautus, the subordinating conjunction sī ‘if’ can be preceded by some or all the other clausal constituents save the verb: saluos domum sī redierō ‘if I shall have returned home safe’, perfidia et peculātus ex urbe et auāritia sī exulant ‘if betrayal and embezzlement and greed are exiled from the city’ (Fortson 2004).

5. According to Clackson (2007), some early IE languages show a clear distinction between the left-detached and fronted position in the sentence. If the sentence includes one of the introductory particles nu, su or ta (sometimes termed S-adverbs), then these normally precede the fronted element.

In these sentences, left-detached nominal phrases are picked up by anaphoric pronouns in the body of the sentence. Enclitics and sentence adverbs are positioned as if the left-detached element was absent; as, Old Irish maisse doíne ní=s toimled (lit. glory of-men, not=of-it he-partook) ‘The glory of men, he did not partake of it’; Lat. N. Pumidius Q.f. [11 other names omitted] heisce magistreis Venerus Iouiae muru aedificandum coirauerunt (lit. Pumidius-nom . . . these-nom magistrates-nom Venus-gen of-Juppiter-gen wall-acc to-be-built-acc supervised) ‘Numerius Pumidius son of Quintus [and 11 others], these magistrates supervised the building of a wall to belong to Venus daughter of Juppiter’.

These examples suggest that the pattern of left-detached elements at the beginning of the sentence should probably be allowed as a possible permutation in the parent language.

6. Right-detached nominal phrases are another common feature of old IE languages. Constituents are placed to the right of the verb as in Gk. oũtin egṑ púmaton édomai metà hoĩs hetároissinnone will I eat last among his comrades’, where the prepositional clause is to the right of the verb édomai ‘eat’. The object has been fronted; the reflexive pronoun hoĩs ‘his own ones’ refers to the topic oũtin more than the grammatical subject egṑ, as a rule (Ramat 1993).

The so-called ‘sentence amplification’ or ‘sentence expansion’, consists in appositional phrases and other adjuncts tacked on to the end of a grammatical sentence (Clackson 2007). The first verse of the Rigveda provides an example: agním īe puróhita yajñáasya devám r̥tvíjam hótāra ratnadhtamam (lit.  Agni-acc I-praise domestic-priest-acc sacrifice-gen god-acc sacrificer-acc invoker-acc best-bestower of treasure-acc) ‘I praise Agni the domestic priest, god of the sacrifice, sacrificer, invoker, best-bestower of treasure’. All of the necessary grammatical information is contained in the first two words, which could stand on their own as a complete sentence, and the sentence is then expanded by the addition of five noun phrases in apposition to the accusative agním.


10.6.4. Wackernagel’s Law and the placement of clitics

1. One of the best known features of clausal syntax of older IE languages is the tendency of unstressed clitic particles to appear second in their clause after the first stressed element, a phenomenon discovered by Jacob Wackernagel in the late nineteenth century.

Examples include (Fortson 2004) Hitt. kiēll=a parnaš ēšar papratar QATAMMA pattenuddu ‘Of this house too may it likewise drive out the bloodshed (and) uncleanliness’; Mycenaean Greek da-mo=de=mi pa-si ko-to-na-o ke-ke-me-na-o o-na-to e-ke-e (lit. people-NOM.=conn=she-ACC. say plot-GEN. communal-GEN. use-ACC. have-INFIN.) But the people say that she has the use of the communal plot’; Vedic tvā mantrā kavisastā vahantu ‘Let the spells recited by the poets lead you hither’; Greek ē̃mos dērigéneia phánē rhododáktulos Ēsbut when early-born, rosy-fingered Dawn appeared’, Lat. tū autem in neruō iam iacēbisBut you will soon be lying in custody, Gothic fram-uh þamma sokida Peilatus fraletan inaAnd at this Pilate sought to release him’.

NOTE. Sometimes, however, one of these clitics appears as the third or fourth word in its clause. Recent research, especially by the American linguist Mark Hale, has shown that Wackernagel’s Law actually involves several processes that usually, but not always, conspire to place unstressed particles in second position in the clause. His discoveries have explained the exceptions to a strict formulation of the law (Fortson 2004).

2. Three types of postpositive clitics (and clitic-positioning rules) can be distinguished (Fortson 2004):

“a. Word-level clitics modify or limit a single word or constituent, and are placed directly after the word or the first element of the constituent. Such clitics tend to have the function of emphasizing the word to which they are attached, or setting it in some kind of contrastive relief or focus (the clitic is boldfaced): Hitt. nu=wa=za apun=pat eši ‘occupy only that (land)’, Vedic pracyāváyanto ácyutā cid ‘the ones who move even unmovable things’.

If the word that such a particle modifies is first in its clause, then the particle appears (coincidentally) second in its clause: Ved. sthir cid ánnāi dayate ví jámbhaieven tough food he cuts apart with his teeth’ (Rig Veda 4.7.10), Lat. hoc quoque maleficium ‘this crime too’.

Such particles, when modifying a phrase, can often come second in the phrase, as in Gk. én ge taĩs Thbais ‘in all of Thebes indeed’.

NOTE.  Some clitics, such as the descendants of PIE -qe, and, can act as word-level clitics as well as sentence connectors.

b. Sentence-connective clitics conjoin or disjoin clauses or sub-clausal constituents. Examples of these clitics are PIE -qe, and, and -we, or. They are attached to the first word of the constituent or clause being conjoined or disjoined, whether that is a single word (Ved. ágna índras ca ‘o Agni and Indra’), a phrase (Lat. silua alta Iouis lūcusue Diānae ‘the high forest of Jupiter or the grove of Diana’), or a clause (Old Avestan yā̊ zī ā̊ŋharə̄ yā̊scā həṇtī yā̊scā mazdā buuaitīindeed (those) who are and who will be, o Mazda’).

NOTE. A more complicated example is explained by Clackson (2007) from “Ved. utá no marcáyād ánāgasa (lit. and part rel-NOM. us-ACC. harm-OPT. innocent-ACC.-Pl.) ‘Or also who would harm innocent us…’ There are two enclitics, the disjunctive particle , which follows the left-detached slot, and the personal pronoun nas following the relative pronoun, which occupies the ‘front’ slot (the pronoun nas appears as no by a process of sandhi).

It should be noted that the position of the particle , which has scope over the whole sentence following the left-detached element, is exactly paralleled by the behaviour of connectives and adverbs with scope over the sentence in Greek, which stand immediately after the left-detached element. Hale (1987a and 1987b) collected evidence for second-position enclitics in Indo-Iranian and showed that, in general, enclitics with scope over the sentence and connectives occurred after left-detached elements, which he refers to as the topic position, whereas enclitic pronouns were placed after the fronted element. Hale claims that the behaviour of these two different sets of enclitics reflects an inherited difference between the two sentence positions.”

 c. Sentential clitics are clitics whose scope is a whole clause or sentence. These include the unstressed personal pronouns as well as a variety of sentential adverbs that serve expressive functions and are often untranslatable into English. They are positioned in various ways. Some are placed after the first stressed word in a sentence and any emphatic or sentence-connective clitics associated with that word, while others (called “special clitics” in the technical literature) are positioned after a particular syntactic structural position in the clause. If the first word in a sentence is a proclitic, that is, an unstressed word that attaches phonologically to a following stressed word, the sentential clitic will of course not come directly after it, as in Gk. eks hēōn gár phāsi kák’ émmenaifor they say that bad things are from us’, where the proclitic eks ‘from’ is not a proper phonological host for the clitic gár.

Sentential clitics occur not infrequently in strings or chains: Ved. v u etán mriyaseindeed you do not die thereby’; Gk. ē̃ rhá moí ti píthoio 'may you indeed now trust me somewhat’. In Vedic utá vā yó no marcáyād ánāgasaor also who would do wrong to us the innocents’, where the pronominal clitic nas (no) ‘us’ is in the second position before the topicalized part, while the disjunctive clitic vā, PIE -w, (and ca, PIE -qe) take the second position within the topicalized part.”

3. There are so-called “preferential hosts”, which are fronted rather than detached, and which are not followed by enclitics in second position, which is apparently a persistent exception to Wackernagel’s Law in languages like Sanskrit, Greek or Latin. Adams (1994) explained it this way:

“Unemphatic pronouns in Classical Latin prose, far from always being placed mechanically in the second position of their colon, are often attracted to particular types of hosts, namely antithetical terms, demonstratives / deictics, adjectives of quantity and size, intensifiers, negatives, temporal adverbs and imperatives. I have suggested that what these hosts have in common is their focused character, and have accordingly argued that enclitic pronouns had tendency to gravitate towards focused constituents. The prominent constituent serving as a host may be at the head of its colon, in which case the clitic will indeed be second, in apparent conformity with Wackernagel’s law. But often the host is in the second or a later position, thereby entailing a place later than second for the pronoun.”

According to Clackson (2007), we should also include relative and interrogative pronouns in this list of preferential hosts for enclitics. “Adams’ findings for Latin also appear to apply well to cases of enclitic pronouns which do not follow Wackernagel’s Law (or even Hale’s modifications of it) in Vedic Sanskrit. So for example in Vedic ágnīomā adyá vām / idá váca saparyáti (lit. Agni-and-Soma-VOC., rel-NOM. today you-two-DAT. this-ACC. speech-ACC. he-praises) ‘Agni and Soma! The one who today hymns you this praise . . .’ Note the placement of enclitic pronoun vām, which is unexplainable in terms of ‘second position’, but can be explained if we consider the temporal adverb adyá ‘today’ as a preferential host.

In other languages too there is evidence for breaches of Wackernagel’s Law, and for the placement of pronominal enclitics after items identified by Adams as preferential hosts. Consider the Greek sentence taken from Krisch (1990) and used by him to support Hale’s arguments: Gk. autàr eg theós eimi, diamperès h se phulássō (lit. conn I-NOM. god-NOM. I-am, thoroughly who-NOM. you-ACC. I-protect) ‘But I am a goddess, the one who protects you steadfastly’. Here the enclitic se follows the relative pronoun h, which comes second in the clause after the emphatically placed adverb diamperés. The relative here can easily be seen as a preferential host, the focussed element in its clause.”

NOTE. Clackson (2007) summarises the situation by establishing that while in Hittite no enclitics are allowed after left-detached elements, or delayed after fronted element, in Sanskrit, Greek and Latin sentence enclitics can stand after left-detached elements, and pronominal enclitics may be delayed, which may therefore reflect the evolution from PIH into Late Indo-European syntax.

10.7. Phrase and Sentence Prosody

1. On the so-called “phonology-syntax interface”, Fortson (2004): “Words belonging to the same constituent that start out as contiguous in the deep structure and stay contiguous throughout the derivation will tend to be grouped together as a single phonological unit, whereas words that only become contiguous through certain kinds of movement sometimes do not. (…) For example, in Greek, clitics normally receive no accent, but if two or more occur in a string, all but the last one get accented, as in ei m tís me theō̃n ‘if no one of the gods me …’. However, in a sequence like doulíāi per en phrení ‘even in bondage in the mind’, there are two clitics in a row but the first is not accented. The reason is that per ‘even’ emphasizes doulíāi ‘in bondage’ and is phonologically attached to it, while en ‘in’ is a preposition that governs phrení ‘the mind’ and is proclitic to that word. The two resultant clitic groups [doulíāi per] and [en phrení] form two separate prosodic groups with what is called a prosodic boundary between them. (A prosodic boundary, incidentally, is not generally audible as a pause or other break). We conclude that the rule placing an accent on the first of two successive clitics applies only if the two clitics belong to the same clitic group.”

2. According to Fortson (2004), noun phrases consisting of a bare noun are much more likely to enter into certain kinds of clitic groups than are noun phrases where the noun is modified by another element. In punctuated Greek inscriptions, interpuncts do not ordinarily separate a definite article from a following noun; but an interpunct is present if the article is followed by a noun modified by another element, indicating a stronger prosodic break between the two.

In Homer, there is a different behaviour of prepositional phrases vs. the positioning of the sentence-connecting conjunction , an enclitic that normally occurs second in its clause. If the clause begins with a prepositional phrase consisting simply of a preposition plus bare noun, the clitic will follow the whole phrase (e.g. eks pántōn dé ‘and of all’…), whereas if the clause begins with a more complex phrase consisting of a preposition followed by an adjective-noun phrase, the clitic will come in between the preposition and the rest of the phrase (e.g. dià khróa kalónand into the fair flesh’…). Therefore, eks pánton is prosodically cohesive enough to function as a single word for the purposes of clitic placement, while dià khróa kalón is not.

A similar phenomenon is found in the Old High German of Notker Labeo: a definite article is written without an accent when preceding a simple noun phrase (e.g. taz héiza fíurthe hot fire’), indicating clisis and destressing of the article, but is written with an accent when preceding more complex noun phrases (e.g díe uuîlsalda állero búrgôthe fortune of all cities’).

3. Through the study of rules and behaviour of poetic meters (metrics), another rule appears (Fortson 2004): “In certain Greek and Roman meters, for example, there is a rule that a sequence of two light syllables in particular verse-positions must belong to the same word. The rule, though, has an interesting exception: a word-break between the two syllables is allowed when one of them belongs to a proclitic (as in the sequence ut opinione ‘that in [his] opinion’). This means that the prosodic group consisting of proclitic plus word was tighter than that consisting of two full-content words – tight enough to behave, for the purposes of the poetic meter, as though there were no word-division.”

10.8. Poetry

1. In the oldest poetry, some common etymologically related Indo-European formulae have been reconstructed; as, klewos dhchitom, immortal fame (cf. Skt. ákitam śrávas, Gk. kléos áphthiton, where the Skt. form is deemed older); mega kléwos, big fame, kléwesa nerōm (<*klewesh2 h2nróm), famous deeds of men, heroes (cf. Gk. kléa andrōn, Ved. śrávas nr̥ṇm); wesu klewos, good fame (cf. Av. vohu sravah, O.Ir. fo chlú). The sun was called ‘the wheel of the sun’, sweljosjo qeqlos, cf. Skt. sū́ryasya cakrás, zaranii.caxra-., (from PII súwarjasja aklas, see Lubotsky’s Indo-Aryan inherited lexicon, IEED, n.d.), Gk. hēlíou kúklos (<PGk sweljohjo quqlos),  O.Ice. sunnu hvél, O.E. sunnan hweogul.

NOTE. Also, the concept “sun chariot” – a mythological representation of the sun riding in a chariot – is typically Indo-European, corresponding with the Indo-European expansion after the domestication of the horse and the use of wheels. Examples include the Trundholm sun chariot of the Nordic deity, Sól, drawn by Arvak and Alsvid (see the cover image); Greek Helios riding in a chariot, Sol Invictus depicted riding a quadriga on the reverse of a Roman coin, and Vedic Surya riding in a chariot drawn by seven horses.

Epithets and adjectives pertaining to the gods might also be found (Beekes 1995); as, dōtōr weswm, those who give goods, riches (cf. Skt. dāt vásūnām, Av. dāta vaŋhvąm, Gk. dōtēres eōn), from wesus, riches, goods, from the same root as ēsús, good.

Other formulae are not etymologically related, but still deemed of PIE origin; as, Skt. pr̥thú śrávas, Gk. kléos eurú, ‘broad fame’; or a common name for the sun Skt. spáśamśvasya jágatas ‘he who spies upon the whole world’ (lit. ‘the moving one’, or ‘the living beings’), similar to Gk. theōn skópon ēde’ kai’ andrōn ‘he who spies upon gods and men’ (Beekes 1995, Clackson 2007).

It is possible to reconstruct formulae from a nexus of correlations, where no single language preserves the complete formula (Clackson 2007). Watkins reconstructs the formula pālāje wīrs pékewaqe (<*peh2- *wīro- *peku-), protect men and livestock, from the correspondence of (etymologically related words underlined) Skt. tryantām asmín grme / gm áśvam púruam paśúm lit. ‘protect in-this village cow, horse, man, (and) flock-animal’ Av. ϑrāϑrāi pasuuā̊ vīraiiā̊ lit. ‘for protection of-cattle (and) of-men’, Lat. pāstōrēs pecuaque salua seruāssīs lit. ‘shepherds farm-animals-and may-you-preserve’, Umb. nerf arsmo uiro pequo castruo frif salua seritu ‘magistrates ordinances men cattle fields fruit safe let-him-preserve’ (Clackson 2007).

2. The metrical structure of Indo-European poetic language was reconstructed by Meillet, although his attribution of the Indo-Greek system to the parent language is not widely accepted today. The oldest Indian and Greek poems were based on a prosodic structure of alternating long and short syllables [see above §2.4.]:

Beekes (1995): “At the end of the line (the cadence) this alternation was strictly regulated, whereas it was free at the beginning. Both the Indic and the Greek systems used the caesura (word-end at a fixed place in the line at a syntactically significant break); in both systems, too, a line had a fixed number of syllables, but a line could also have a variant with a syllable less (catalexis). [Meillet] saw an exact similarity between the eleven syllable line used by the Greek poetess Sappho and the triubh of the Rigveda ( ¯ long, ˘ short, x long or short, | caesura; a begins the cadence);

triubh  x x x x | x ˘ ˘, ¯ ˘ ¯ x

Sappho  ¯ ˘ ¯ x | ¯ ˘  ˘ ¯ ˘ ¯ x

Meillet’s theory was further developed by Roman Jakobson, who believed that the oldest Slavic metric system was comparable to the one Meillet described, and derived from an Indo-European decasyllabic line. A similar continuation could be seen in a verse-tipe utilized in Old Irish.”

It is probable then that the Proto-Indo-European verse consisted of a fixed number of syllables, as was the case with Avestan. Longer lines would have contained a caesura (division between phrases).

Beekes (1995) takes as example of PIE poetry a stanza from a hymn of Zarathustra in Gatha-Avestan (Yasna 44.4), with a phonetic transcription: The line consists of four plus seven syllables (with a caesura after the fourth syllable):

Tat ϑvā pr̥sā

Kas-nā dr̥ta


Kah vaʔatāi

Kas-nā vahauš,


This I ask Thee,

Who has upheld

from falling down

Who to the wind

Who, Wise One, is

r̥š mai vauca, Ahura:

zam ca adah nabās ca

kah apah urvarās ca?

vanmabyas ca yaugi āsuu?

Mazdā, dāmiš manahah?


tell me truthfully, O Lord:

the earth below and heavens [above]

who the waters and the plants?

and the clouds has yoked the swift [horses]?

the founder of Good Thinking?

Beekes (1995) continues: “It is probable that this kind of song is of Indo-European origin, for we find a parallel to it in the Edda (Alvíssmál 15):

Segðu mér, Alvíss,

voromc dvergr, et vitir –

hvé sú sól heitir,

heimi hveriom í.


Tell me this, Alwis

I think, dwarf, that thou knowest

What the sun is called,

by all creatures of the world.

ǫll of rǫc fira


er siá alda synir,



of everything in the world


the sun which people see,

In the answer the sun is referred to, among other things, as a ‘wheel’, hvél [ see above]. The Germanic verse here is stress verse, and is not related in this respect to Indo-European poetry.”

10.9. Names of persons

4.8.1. The use of two-word compounds for personal names was common in PIE; as, Suklewos, of good fame, with cognates found in poetic diction, cf. Gk. Eukles, and Skt. Suśráva-, or Illyr. Vescleves-.

NOTE. The use of two-word compound words for personal names is common in IE languages (see above §10.4.2). They are found e.g. in Ger. Alf-red ‘elf-counsel’, O.H.G. Hlude-rīch ‘rich in glory’, O.Eng. God-gifu ‘gift of God’ (Eng. Godiva), Gaul. Orgeto-rix ‘king who harms’, Gaul. Dumno-rix ‘king of the world’, Gaul. Epo-pennus ‘horse’s head’, O.Ir. Cin-néide (Eng. Kennedy) ‘ugly head’, O.Ind. Asva-ghosa ‘tamer of horses’, O.Ind. Asvá-medhas ‘who has done the horse sacrifice’, O.Pers. Xša-yāršā (Gk. Xérxēs) ‘ruler of heroes’, O.Pers. Arta-xšacā ‘whose reign is through truth/law’, Gk. Sō-krátēs ‘good ruler’, Gk. Mene-ptólemos ‘who faces war’, Gk. Hipp-archus ‘horse master’, Gk. Cleo-patra, Pátro-klos ‘from famous lineage’, Gk. Arkhé-laos ‘who governs the people’, O.Sla. Bogu-milŭ ‘loved by god’, Sla. Vladi-mir ‘peaceful ruler’, from volodi-mirom ‘possess the world’; etc. Individual names may further be modified through the use of suffixes to form hypocorisms.

4.8.2. Other area in which it is suspected the retention of ancient Proto-Indo-European personal names is the use of animal names or numerals, composed of one stem; as Wqos, wolf, cf. O.Ir. Olc, O.Eng. Wulf, Gk. Lukos, Skt. Vŕka; or, Qétwtos, fourth, cf. Lat. Quārta, Lith. Keturai, Russ. Četvertoj, Gk. Tetartíōn.

NOTE. The word for ‘name’ and possible Indo-European names can be found in Beekes (1987), Markey (1981), Pinault (1982), Schmitt (1973), and Watkins (1970).

4.8.3. Further, the syntactical indication of the father’s name also dates from Proto-Indo-European, whether by adding the name of the father in the genitive, in the sense of ‘son of X’, or by adding a possessive adjective that is derived from the name of the father.

NOTE. An example of the former is Hadubrand Heribrandes suno; an example of the later is Myc. a-re-ku-tu-ru-wo e-te-wo-ke-re-we-i-jo, i.e. Alektruwōn Etewoklewehijos ‘Alektruwōn, son of Etewoklewēs’, or Russ. Nikolaj Sergejevich. Patronymics ending in -ios (later -ius) led to what is called the nomine gentile in Rome, cf. Gaius Iulius Caesar with Gaius = praenomen < individual name, Iulius = nomen gentile < patronymic and Caesar = cognomen.

4.8.4. When considering the giving of names to individuals, one departs generally from the basis of the free men.

Whereas the man is addressed using the individual name, a simple ‘oh woman’ suffices in the case of woman. “The woman is treated more as a typus, the man as an individual”. Wackernagel (1969) makes clear that the same forms of address were adopted for interactions with the gods.

According to Meier-Brügger (2003), to say that the Indo-Europeans were not very different from the Romans and Greeks would not likely be too far from the mark. In Rome, women generally carried only the nomen gentile, cf. Cornelia, Julia, etc. In the case of the Greeks, most names of women are simply feminine forms of masculine names of individuals, e.g. Myc. a-re-ka-sa-da-ra, i.e Aleksandrā (corresponding to Aleks-anōr ‘who fights off men’), Hom. Andromákhē, from Andrómakhos ‘who fights with men’, etc.






Part IV

Texts & Dictionary








By Fernando López-Menchero