The University of Texas at Austin; College of Liberal Arts
Hans C. Boas, Director :: PCL 5.556, 1 University Station S5490 :: Austin, TX 78712 :: 512-471-4566
LRC Links: Home | About | Books Online | EIEOL | IE Doc. Center | IE Lexicon | IE Maps | IE Texts | Pub. Indices | SiteMap

Proto-Indo-European Syntax

Winfred P. Lehmann

2. The Syntax of Simple Sentences

2.1. The OV Structure of PIE as Demonstrated by Vedic.

Linguists have generally maintained that the sentence is the basic unit of language (Brugmann 1904a:623, Paul 1920:121). Linguistic analysis, whether of attested or reconstructed languages, must therefore concern itself in the first instance with the sentence. This chapter examines the forms of the simple sentence in PIE. It will be followed by chapters dealing with modifiers of its constituents, with syntactic categories, and with selected lexical entries.

As noted in Chapter 1, the fundamental order of sentences in PIE is OV. Support for this assumption is evident in the oldest texts of the materials attested earliest in the IE dialects. 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, as well as the other patterns discussed in § 1.4. All of these constructions, as well as the basic sentence pattern, provide evidence that PIE was OV in structure.

Berthold Delbrück, the great syntactician of IE, assembled the pertinent data for Vedic Sanskrit and arrived at conclusions about these syntactic patterns which would lead to the characterization of PIE as OV. But he did not discuss the interrelationships between the individual syntactic constructions or the importance of fundamental sentence order, whether VO or OV. Yet, since the data for Vedic Sanskrit were given by him, evidence in Indic for the word order of PIE will be cited first.

2.1.1. Basic Sentence Pattern and Further Verbal Constructions (Delbrück 1888:17).

1. Śatapathabrāhmaṇa (ŚB) víśaḥ kṣatríyāya balíṃ haranti
villagers to-prince tax they-pay
‘The villagers pay taxes to the prince.’

Comparative constructions, with the pattern standard (‘ghee’)-pivot (ablative)-comparative (‘sweeter’) (Delbrück 1888:113):

2. RV 8.24.20. ghṛtā́t svā́dīyaḥ
ghee-from sweeter
‘sweeter than ghee’

In many of Delbrück's examples this construction is breaking down, for they exhibit the order comparative-pivot-standard, as in:

3. Taittirīya Saṁhitā (TS) pā́pīyān áśvād gardabháḥ
lesser horse-from donkey
‘A donkey is less valuable than a horse.’

But even though the order typical for OV languages is being abandoned, there are few examples in Vedic Sanskrit of a comparative construction expressing the pivot with a particle, as in VO languages like English (Delbrück 1888:196). Accordingly, evidence in Indic suggests that the inherited pattern of the comparative construction was OV. This assumption is supported by evidence from Hittite, Greek, and other dialects which will be given below.

Postpositions Rather than Prepositions. Delbrück states as a general principle that the genuine prepositions “traditionally stand behind the case form they govern” (1888:21). His section on these (1888:440-470) contains many examples, of which the first is:

4. ŚB yád imā́n̐l lokā́n áti caturthám ásti
what these worlds beyond fourth it-is
‘What is fourth beyond these [three] worlds.’

The evidence in Vedic for postpositions accordingly supports strongly the conclusion that the word order of PIE was OV.

2.1.2. Evidence from Nominal Modifying Constructions, Such as Relative Clauses.

Since relative clauses will be discussed at length in the following chapter, Delbrück's conclusions may simply be noted here about the position of the relative clause with regard to its antecedent in Vedic. Both in poetry and in prose relative clauses commonly precede their antecedent when it is (Delbrück 1888:559, 564). John Avery (1881) found that relative clauses preceded main clauses in more than 50 percent of the “nearly 4,000” passages in the Rigveda.

5. RV 2.13.2. yás tā́kṛṇoḥ prathamáṃ sā́sy ukthyàḥ
who these-things-you-did first, that-you-are to-be-praised
‘You who did these things first are to be praised.’

(The literal translation given here is clumsy; actually the sa introducing the second clause is a reflex of a sentence-introductory particle rather than a pronoun, as we came to know when Hittite was analyzed.) Another of Delbrück's examples shows the same construction in the plural:

6. Aitareya Braḥmāna (AB) 3.34.1. ye ’n̄gārā āsaṅs te ’n̄giraso ’bhavan
who coals were these Angiras became
‘Those who were coals became Angiras.’

As might be expected from the position of relative clauses, descriptive adjectives also precede their nouns (Delbrück 1888:19-20):

7. mānuṣī́ vā́k
  ‘human voice’

Attributive genitives also precede the nouns they modify (Delbrück 1888:19-20):

8. devā́nāṃ hótā
  of-the-gods priest
  ‘priest of the gods’

Nominal modifying constructions accordingly support the conclusions concerning OV word order.

2.1.3. Evidence from Absence of Prefixation, from Conjunctions, and from Gapping.

Three further characteristics of arrangement described by earlier scholars provide evidence of OV word order. The first is the absence of prefixes. Antoine Meillet states flatly that PIE did not contain prefixes (1937:151). Delbrück concludes that “prepositions” are not combined with verbs (1888:44-51). For often they follow the verb; but when they precede it, other words may stand between the “preposition” and the verb.

9. TS ápa támaḥ pāpmā́naṃ hate
away darkness evil she-drives
‘She drives off darkness and sin.’

Vedic accordingly provides evidence that there were no prefixes in PIE, in keeping with Meillet's conclusion.

Conjunctions follow rather than precede the second or further word of a conjoined series (Delbrück 1888:22). For the most frequent of these, ca, Hermann Grassmann gives many examples, as Delbrück indicates (1888:474).

10. RV 1.12.10. yajñáṃ havís ca
sacrifice libation and
‘sacrifice and libation’

Gapping. Patterns of gapping still require much detailed investigation. But Vedic contains examples of OV gapping, in which the identical verb is deleted from the first of two conjoined clauses and maintained in the second, as in the following example (Ptc. = particle):

11. RV 1.102.7.
út te śatā́n maghavann úc ca bhū́yasa
Ptc. your 100-from O-blessed Ptc. and more
út sahásrād ririce kṛṣṭíṣu śrávaḥ
Ptc. 1000-from exceeds among-men glory
  ‘Your glory, O blessed one, exceeds more than a hundred, even more than a thousand among men.’

The two clauses conjoined by ca, introduced by úd (út, úc), have the verb deleted in the first.

On the basis of the above-listed characteristics of OV structure, it is clear that Vedic Sanskrit was largely OV, even though VO patterns were being introduced. Moreover, the Vedic evidence strongly supports the conclusion that the earlier language was essentially OV.

2.1.4. Order of Nominal Sentences.

In accordance with the OV structure of sentences containing objects and verbs, sentences containing predicate substantives may be expected to have the order topic-predicate substantive. This is the order in Hittite:

12. J. Friedrich 1960:117-118. attaš aššuš
father good
‘The father [is] good.’

13. J. Friedrich 1959:72, § 42. 1 GÍN KUBABBAR kuššan šet
1 shekel silver reward his
‘One shekel of silver [is] his reward.’

As Friedrich points out, when such sentences were in the past tense, the copula could not be omitted and had to be placed at the end of the sentence. For this reason I now consider erroneous my earlier assumption that the Sanskrit order N ... Pred N was marked (Lehmann 1969: 10-11). For Vedic as well as Hittite and PIE, the unmarked order must have the predicate noun or adjective in final position, as in:

14. RV 8.48.15. tvám naḥ soma viśváto vayodhā́s
thou to-us soma on-all-sides strength-giver
‘Thou, soma, should be our strength-giver on all sides.’

Predicate noun and adjective sentences are accordingly parallel with sentences containing verbs in having the “verbal” element at the end of the sentence. They may also be expanded with optional substantival phrases and in this way show expansions of the types discussed below.

2.2. Evidence from Hittite for OV Structure of PIE.

In support of the assumption that OV structure must be assumed for PIE, material will be cited from Hittite, with references to similar patterns in other dialects.

For Hittite the examples included here contain evidence for six selected patterns of OV order, without having a special sentence exemplifying each, for each example contains two or more of the patterns.

Verb-final order and order of adjectives with regard to their nouns are illustrated in the following sentence (Otten and Souček 1969:18, § 14):

15. DUMU.É.GAL šuppi watar parā epzi
  nobleman pure water preverb he-brings
  ‘The nobleman brings pure water.’

A proposed genitive, anzēl, and the OV comparative pattern with TI-anni as standard are represented in the following example (J. Friedrich 1960: 127):

16. namma- kan anzēl TI-anni UL ŠA BELU-NI TI-tar nakki
  moreover Ptc. of-us life (Dat.-Loc.) not Gen.-Ptc. lord-our life important
  ‘Moreover, (if) the life of our lord is not more important than our life...’

An instance of a postposition, piran, and of a preposed relative clause is found in the following example; this example also illustrates the placement of verbal complements before the finite verb akuwanna (Raman 1973:58):

man LUG[A]Lwaš piran šeškanzi kuiš hazzizzi
when king before they-sleep who he-acts-appropriately
nu- šše GEŠTIN-an akuwanna pianzi
  Ptc. to-him wine to-drink they-give
  ‘When they sleep before the king, then people give wine to the one who acts appropriately.’

Accordingly, nominal modifiers, that is, relative constructions, genitives, and adjectives are consistently placed before nouns; the postpositional constructions and the order of comparative constructions and clause order are also OV. Hittite materials accordingly support the conclusion that PIE was OV.

2.3. Evidence from Other Dialects for OV Structure of PIE.

The other early dialects also support these conclusions about OV order; examples will be given here primarily from Greek, with supplementary materials in other dialects.

The basic word order of simple sentences is OV in early Greek (Schwyzer 1950:693-696), as illustrated in the first clause of the Odyssey:

18. Odyssey 1.1. ándra moi énnepe
man to-me you-tell-of
‘Tell me about a man...’

Early Latin also has OV order (Leumann 1965:397-398), as in an example from Archaic Inscriptions (Warmington 1959:154):

19. Honce loucom nequs violatod
  this grove no-one he-should-violate
  ‘No one should violate this grove.’

In comparative constructions the standard is indicated by the genitive in the oldest Greek texts, and this commonly precedes the adjective (Schwyzer 1950:98-101). In this use the genitive replaces the ablative:

20. Iliad 16.722.
aíth’, hóson hḗssōn eimí,
Oh, to-what-extent weaker I-am
tóson séo phérteros eíēn
to-that-extent from-you stronger I-might-be
  ‘I wish I were as much stronger than you as I am weaker.’

For Latin an example of the preceding standard in the ablative may be illustrated with quō in the following quotation (Leumann 1965:136, 162-163.)

21. Vergil, Aeneid 1.544.
Aenēās, quō iūstior alter
Aeneas from-whom more-perfect another
nec pietāte fuit  
not-ever in-piety he-was  
  ‘Aeneas, than whom no other was ever more perfect in his piety.’

While case forms were commonly used in the earliest Greek texts to indicate nominal relationships in the sentence, they could be supplemented with particles, which subsequently developed into prepositions. As in Hittite and Vedic, these often followed their nouns, and accordingly they must be viewed as postpositions in OV languages (Schwyzer 1950:417-421).

22. Iliad 11.50. ásbestos boḕ génet’ ēȏthi pró
unquenchable but cry it-arose Dawn before
‘But an unquenchable cry arose before Dawn.’

In Latin this position of “prepositions” was so frequent that it was given a special name, anastrophe (Leumann 1965:214-217, 692-693). It is particularly notable in frozen expressions, such as mēcum ‘with me’, sēcum ‘with himself’.

Moreover, prepositional adverbs were used in conjunction with verbs, but not as prefixes (Schwyzer 1950:419-424; Leumann 1965:214-215). Many syntactic patterns subsequently maintained in the dialects give evidence of this situation, such as the use of “separable prefixes” in German today: aufführen ‘produce’ but er führt...auf ‘he produces’, and the placing of particles between the apparent verbal prefix and the verb root in some Gothic forms, such as ga-u-laubjats ‘do you two believe’ (see Streitberg 1920:161). Another such pattern is the use of a prefix with only the first of two verbs, as in Xenophon, Cyropaedia 7.1.1.:

23. prosḗnegkan hoi therápontes empieȋn kaì phageȋn (not emphageȋn)
  ‘The servants brought [things] to drink and to eat.’

As an OV language, PIE did not have prefixes. The prefixes introduced when the dialects became VO were not integrated into them for some time, as these examples demonstrate.

Relative constructions in an OV language may have finite or nonfinite forms preposed to nouns. Japanese formerly had a special attributive verb form which was used before nouns and particles, but in contemporary standard Japanese the normal finite forms are used attributively to provide relative constructions. In Turkish, on the other hand, special forms are used in relative constructions and are called participles. When the IE dialects were becoming VO, only this participial construction was found in place of relative constructions, as in Greek:

24. Odyssey 9.259. hēmeȋs toi Troíēthen apoplagkhthéntes Akhaioì
we indeed from-Troy returning-back Achaeans
‘We are Achaeans who are returning from Troy.’

As is well-known, such participial constructions could be placed either before or after the nouns they modified, as in the following example:

25. Iliad 24.3. autàr Akhilleùs
but Achilles
klaȋe phílou hetárou memnēménos 
he-wept dear friend remembering  
  ‘But Achilles wept, remembering his friend.’

Such participial constructions came to be widely used in the various IE dialects and eventually were not associated with relative constructions. Yet in early materials they cannot simply be treated as adjectival participles, as in the following example cited by Schwyzer (1950:392, also 618-619):

26. Odyssey 5.6. méle gár hoi eṑn en dṓmasi númphēs
it-worried for to-her being in house of-the-goddess
  ‘For he who was in the house of the goddess was a concern to her.’

In this example the participle eṑn is equivalent to a finite form of a clause, which generally was introduced by the relative pronoun hós in Greek.

The assumption that participles, like those in the examples cited above, were developments of PIE relative constructions with finite verbs may be supported by recalling an early use of adjectives which has long intrigued linguists (Wackernagel 1928:65-75). In this use the adjectives function as substantives modified by a relative construction:

27. Odyssey 14.257. pemptaȋoi d’Aígupton eürreítēn hikómestha
as-fivers Egypt beautifully-watered we-came
  ‘We who had traveled five days came to beautifully watered Egypt.’

Wackernagel gives further examples in his excellent discussion, in which he also indicates the puzzled attempts of previous scholars to explain this use (1928:67-68). When the adjectives are analyzed as reduced relative constructions, the use becomes clear. As Wackernagel indicates, the construction has been maintained to various periods of the dialects and is still characteristic of Slavic.

This adjectival construction in the early dialects also illuminates the view that descriptive genitives are derived from relative constructions. For many nominal modifiers are still expressed adjectivally in the early period, when later they are expressed through genitives. Again Wackernagel cites many examples (1928:68-74), such as Nestoréē nēȗs ‘Nestorian ship = Nestor's ship’ and Paelignian iouiois puclois ‘to Jovian boys = to the boys of Jupiter’. The parallelism between the adjectival and genitival attributes is nicely illustrated by one of Wackernagel's well-chosen quotations (1928:72):

28. Iliad 2.54. Nestoréēi parà nēï̀ Puloigenéos basilȇos
Nestorian beside ship of-Pylian-race of-king
‘beside the ship of Nestor, the Pylian king’

These examples may illustrate the attributive constructions which developed in the early dialects from OV relative constructions: on the one hand, descriptive adjectives; on the other, genitives. As Wackernagel also indicates, they are further related to compounds and to patronymics (see also § 3.7).

The older pattern, of compound adjective as well as patronymic, is exemplified in:

29. Odyssey 1.30. tón rh’ Agamemnonídēs tēleklutòs éktan’ Oréstēs
him Ptc. Agamemnonian far-famed slew Orestes
‘Him Orestes, the far-famed son of Agamemnon, killed.’

The older patronymic is combined with the younger genitive construction in:

30. Odyssey 1.429. Eurúklei’, Ṓpos thugátēr Peisēnorídao
Eurycleia of-Ops daughter of-Peisenorid
‘Eurycleia, daughter of Ops, the son of Peisenor.’

In time the genitive construction came to be more frequent than the adjective, which as Wackernagel states (1928:74) has come to be restricted to indicating characteristics in some dialects and no longer is used to indicate relationships.

The selected syntactic patterns discussed here illustrate that many syntactic features of PIE and the dialects can be accounted for by assuming that PIE was OV in structure and that it was developing to a VO language at the time of the early dialects. Many OV characteristics are found in the IE dialects spoken today, even those which are basically VO; such are the adjective-noun order of English and this order for a restricted number of French adjectives. The description of such patterns in their change through five millennia of history of each dialect is a task for the specialists in these dialects. A unidirectional change cannot simply be assumed, for external influences may have reintroduced OV patterns in VO languages (Lehmann 1970a:286-305). But on the basis of patterns such as those cited from the early dialects, PIE can be assumed to be OV. Its various sentence patterns are examined below.

2.4. The Structure of the Simple Sentence in PIE.

When the simple sentence is discussed in studies of PIE syntax it is generally assumed that the minimal sentence consists of two elements: a subject and a predicate. Moreover, in keeping with this assumption, sentences which consist of one element are considered aberrant. The so-called impersonal sentences, like those referring to natural events, e.g., Lat. pluit ‘It's raining’, are treated as fragments. At best, the second singular imperative is recognized as a complete sentence. The point of view which led to these conclusions is based on experience restricted to Indo-European languages. The scholars who characterized one-word sentences in this way did not take OV languages into account. In OV languages a sentence is characterized by a verb, accompanied or not by a substantive; if there is a substantive, it is in the first instance an object rather than a subject. Accordingly, Indo-Europeanists and linguists in general must modify their views of sentence structures in language.

In conversation between two friends the following statements are natural in Japanese.

31. Taroo o yonda
  Taroo Obj. Ptc. called (past of yobu ‘call’)
  ‘I called Taroo.’

32. Taroo o yonda ka (ka = Int. Ptc.)
  ‘Did you call Taroo?’

33. Wakatta
  ‘I understand’ < ‘have perceived’.

All of these sentences, like many sentences in OV languages, lack explicit subjects. The simple sentence of PIE may then be expected to consist of a verb and, depending on the lexical characteristics of the verb, one or more objects in addition. A subject, like expressions of time, place, and manner, would be optional, not mandatory. But for transitive verb forms an object would be mandatory. It is the influence of Western logic, based on the sentence patterns of Greek when it was already largely SVO, that has led linguists to assume for the simple sentence pattern of language a structure consisting of subject and predicate, whether expressed in this way or by the base rule: S NP VP. Investigation of VSO languages as well as of (S)OV languages leads to the conclusion that the subject is far less central in the sentence than is the object, without taking into consideration the verb, which is the essential element, as we have noted in Chapter 1.

The so-called impersonal sentences accordingly exemplify the simplest form of PIE sentences. Among the best-known examples are those referring to natural events, such as:

34. Skt. várṣati, Lat. pluit, Goth. rigneiþ ‘it is raining’

(For others see Brugmann 1925:22-24). As is also well known, 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. One result of this purely linguistic requirement has been a great deal of speculation on the antecedent of it in such constructions. The speculation is largely pointless; it was 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:

35. RV 10.34.11. kitaváṃ tatāpa
with-regard-to-the-gambler there-is-pain
‘It pains the gambler.’

Latin verbs may also be compared, such as:

36. miseret, pudet, taedet
  ‘It makes one pitiful, ashamed, bored.’

These verbs may be accompanied by one or more nouns, as in Cicero's Oration for Milo 34, 92 (Hale-Buck 1903:186-187):

37. eōrum nōs miseret
  of-them us it-makes-pitiful/there-is-pity
  ‘We feel pity for them.’

The genitives in such impersonal sentences, like the accusatives, have meanings which are determined by the functions of these cases, and they cannot be parsed as “governed” by the impersonal verb. Presumably because of this situation Meillet included in his definition of sentences of PIE the characterization that every word was autonomous (1937:355). In keeping with this observation of Meillet's, examples of sentences like that cited from Cicero may be interpreted as typical constructions in a language which is OV, even though in Cicero's time this pattern was archaic. In PIE, verbs could be supplemented by one or more nominal elements, in accordance with the lexical properties of the verbs in question. Evidence for this assumption is found in dialects as late as Old High German, as in the following example (Erdmann 1882:167):

38. Otfrid 4.5.44. thes gánges thih nirthrúzzi
of-the way you may-there-not-be-weariness
‘Do not let yourself be wearied of the way’

In this sentence the accusative thih cannot be taken as object of the impersonal verb irthriazan, but, like the genitive ganges, must be interpreted on the basis of the inherent meaning of the case.

Because sentences consist mandatorily of verbs, which may or may not be accompanied by nouns, the P rule for the proposition is written with the underlying nominal constituents as potential:

Prop. V (Agent) etc.

The nominal constituent accompanying any verb will be indicated in the lexicon, e.g.:

*peγ- ‘drink’ _____ Agent Target (Time) (Place) (Means)

As this lexical entry indicates, the possible further constituents are included in parentheses; lexical entries for selected roots are given in Chapter 6. The rules for surface expression of the underlying nominal constituents are also given there.

In PIE sentences various case forms could be used with verbs. Examples are given below, first for mandatory patterns and then for sentences in which the nominal elements are optional (see also Brugmann 1911:494-651).

Some verbs are accompanied by nouns in the nominative case, as is as- ‘be’, in:

39. RV 8.48.8. táva smasi vratíās
your we-are devotees
‘We are your devotees.’

Others are accompanied by the accusative, in constructions labeled “direct object,” as is rudh- ‘withhold’ in:

40. RV 10.34.12. dhanā́ ruṇadhmi
not money I-withhold
‘I do not withhold money.’

Compare Hittite nu hurtialan harmi ‘beaker I-hold’ ‘l hold a beaker’ (Otten and Souček 1969:18, § 7). Others are accompanied by nouns in the instrumental, as is div- ‘play’ in:

41. RV 10.34.13. akṣáir mā́ dīvyaḥ
with-dice do-not you-play
‘Do not play (with) dice.’

In the dative:

42. RV 10.34.14. mṛlátā no
‘Be gracious to us.’

In the ablative:

43. RV 10.34.5. parāyádbhyo áva hīye sákhibhyaḥ
from-departing behind I-am-left from-friends
‘I am left behind by the departing friends.’

In the genitive:

44. RV 8.48.8. tásya viddhi
‘Know that.’

In the locative:

45. RV 10.34.13. vitté ramasva
‘In property rejoice.’

As illustrated by these examples, which are complete sentences in the passages from which they are taken, Vedic simple 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. 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 was in no way 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 *deγ- had in its simplest sense the meaning ‘present’ and was often unaccompanied by any nominal expression, as in:

46. RV 2.30.7. dádad
who he-gives
‘who makes presents’

In the same way, other roots which may seem to require more than one noun to accompany them are found in their simplest constructions with a simple noun, if they have any complement whatsoever.

When subjects are explicitly expressed, the nominative is the case employed, as in:

47. RV 10.34.1. prāvepā́ bṛható mādayanti
dangling-ones me of-the-high-tree they-please
‘The dice (consisting of the nuts of the vibhī́daka tree) please me.’

Expression of the subject is the most prominent extension of simple sentences to include more than one substantival expression.

Besides 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. Such constructions must be distinguished from the inclusion of additional nouns whose case forms indicate adverbial use. Since the use of cases in the dialects and in PIE is among the best-described features of their syntax, the various uses of cases will not be discussed here (see for example Krahe 1972:56-108). For our conern is the set of basic sentence patterns.

Few verbs are mandatorily accompanied by two nouns. Of such constructions the most frequent with simple verbs is the use of the dative in addition to the accusative, as in:

48. RV 10.14.11. tā́bhiām enaṃ pári dehi
to-those-two him over you-give
‘Give him over to those two.’

Yet even in such sentences the additional case forms are not mandatory for the verb and accordingly are not essential in the sentence. Rather, the additional substantive may be viewed as supplemental or adverbial. This problem is found also for other cases than the dative, such as the instrumental and ablative as in the following examples:

49. RV 1.32.5. áhan vṛtrám ... índro vájreṇa
he-killed Vṛtra ... Indra with-bolt
‘Indra killed ... Vṛtra with his bolt.’

50. RV 7.5.6. tváṃ dásyūm̐r ókaso agna ājaḥ
you enemies from-house O Agni you-drove
‘You drove the enemies from the house, O Agni.’

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. In view of such uses, all instances of more than one noun in sentences may be taken as facultative or a result of embedding or of other complex constructions. A further example is the causative accompanied by two accusatives:

51. RV 2.37.6. devā́n̐ uśataḥ pāyayā havíḥ
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 (1916:305), in a corresponding simple sentence this noun would have been given in the nominative:

52. devā́ havíḥ pibanti
  gods libation they-drink
  ‘The 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. Complementary elements are introduced by means of embedding, as noted in greater detail below.

2.5. Simple Sentences Accompanied by Two or More Substantives.

Nonmandatory case forms are found in great variety, as may be determined from the studies of substantival inflections and their uses (Delbrück 1893:173-400, Brugmann 1911:464-651, Macdonell 1916:298-328, and many others). Such extended sentences will be exemplified below with five groups of adverbial elements: (1) circumstance, purpose, or result; (2) time; (3) place; (4) manner; (5) means.

2.5.1. Circumstance, Purpose, or Result.

Additional case forms may be used to indicate the purpose, result, or circumstance of an action, as is the instrumental in the following sentence:

53. RV 8.48.8. mṛḷáyā naḥ suastí
be-gracious to-us for-well-being
‘Be gracious to us for our well-being.’

The dative was commonly used in this sense, as in the widely discussed infinitival forms:

54. RV 8.48.4. prá ṇa ā́yur jīváse soma tārīḥ
ahead our years for-living soma you-cross
  ‘Extend our years, soma, for our living [so that we may live long].’

Compare the Hittite dative noun form haluki in the example below (Raman 1973:106):

nu-kan mNana-LUin kuin DUMU.LUGAL ANA mNuwanza
Ptc.-Ptc. NanaLUiš whom prince to Nuwanza
haluki para nehhun
for-the-purpose-of-the-embassy away I-sent
  ‘and the prince NanaLUiš whom I sent to Nuwanza to convey the message’

When an animate noun is involved, this use of the dative has been labeled the indirect object:

56. RV 7.71.1. riṇákti kṛṣṇī́r aṛuṣā́ya pánthām
he-yields black to-ruddy path
‘Black night gives up the path to the red sun.’

As these examples may indicate, the dative, like the other cases, must be interpreted with reference to the lexical properties of the verbal element.

2.5.2. Time.

A further adverbial segment in sentences indicates the time of occurrence. The cases in question are various, as the following example may indicate:

57. RV 7.71.1. dívā náktaṃ śárum asmád yuyotam
by-day during-night arrow from-us ward-off
‘By day and during the night protect us from the arrow.’

The nominal form dívā, which with change of accent is no longer an instrumental but an adverbial form outside the paradigm, and the accusative náktaṃ differ in meaning. The instrumental, like the locative, refers to a point in time, though the “point” may be extended; the accusative, to an extent of time. Differing cases accordingly provide different meanings for nouns marked for the lexical category time.

2.5.3. Place.

Nouns indicating place also differ in meaning according to case form. The accusative indicates the goal of an action, as in Lat. Rōmam īre ‘go to Rome’, Hitt. tuš alkištan tarnahhe ‘and those (birds) I release to the branch’ (Otten and Souček 1969:38 § 37). The instrumental indicates the place “over which an action extends” (Macdonell 1916: 306): sárasvatyā yānti ‘they go along the Sarasvatī’. The ablative indicates the starting point of the action: sá ráthāt papāta ‘he fell from his chariot’; and the following example from Hittite (Otten and Souček 1969:20):

58. iššaz-(š)mit lālan AN.BARaš [d]āi
  from-mouth-their tongue of-iron he-takes
  ‘He takes the iron tongue out of their mouths.’

The locative indicates a point in space, e.g., Skt. diví ‘in heaven’ or the locative kardi in the following Hittite example (Otten and Souček 18, § 12):

59. kardi- šmi- i̯a- at- kán dahhun
  in-heart your- and- that- Ptc. I-took
  ‘And I took away that [illness which was] in your heart.’

In addition, Hittite has a case form ending in a which grammars label the dative (J. Friedrich 1960:44) but which Heinrich Otten and Vladimir Souček call the directive (1969:62-63; 30, § 8):

60. teššummiuš- a takna hariemi
  beakers and in-earth I-bury
  ‘And the beakers I bury in the earth.’

Nouns with lexical features for place and for time may be used in the same sentence, as in the following example:

61. RV 10.34.10. ástam úpa náktam eti
house to night he-goes
‘He goes during the night to the house.’

Although both nouns are in the accusative, the differing lexical features lead to different interpretations of the case. In this way, inflectional markers combine with lexical features to yield a wide variety of adverbial elements.

2.5.4. Manner.

Among the adverbial elements which are most diverse in surface forms are those referring to manner. Various cases are used, as follows. The accusative is especially frequent with adjectives, such as Skt. kṣiprám ‘quickly’, bahú ‘greatly’, nyák ‘downward’. The instrumental is also used, in the plural, as in Skt. máhobhiḥ ‘mightily’, as well as in the singular, sáhasā ‘suddenly’.

Similar to the expression of manner is the instrumental used to express the sense of accompaniment:

62. RV 1.1.5. devó devébhir ā́ gamat
god with-gods hither he-may-come
  ‘May the god come [in such a way that he is] accompanied by the other gods.’

This use of the instrumental is particularly prominent with nonanimate nouns such as javas ‘haste’:

63. RV 4.22.6.
ádhā ha tvád vṛṣamaṇo bhiyānā́ḥ
there indeed of-you strong-hearted fearful
prá síndhavo jávasā cakramanta
ahead streams with-haste they-strode
  ‘There indeed fearing you, O strong-hearted, the streams strode onward hastily.’

The ablative is also used to express manner in connection with a restricted number of verbs such as those expressing ‘fear’:

64. RV 7.21.3. réjante víśvā kṛtrímāṇi bhīṣā́
tremble all made-ones through-fear
‘All creatures tremble fearfully.’

2.5.5. Means.

Adverbial expressions of means are expressed especially by the instrumental:

65. RV 1.32.5. áhan vṛtrám ... índro vájreṇa
he-killed Vṛtra ... Indra with-bolt
‘Indra killed ... Vṛtra with his bolt.’

As in this example, the noun involved frequently refers to an instrument. Compare also the Hittite example (Otten and Souček 38, § 26):

66. ta kalulupuš šmuš gapinit hulaliemi
  Ptc. fingers their with-thread I-wind
  ‘I wind the thread around their fingers.’

Animate nouns may also be so used. When they are, they indicate the agent:

67. RV 1.36.18.
agnínā turváṣaṃ yáduṃ parāváta
through-Agni Turvaṣa Yadu from-far
ugrā́devaṃ havāmahe
Ugradeva we-call
  ‘Through Agni we call from far Turvasa, Yadu, and Ugradeva.’

This use led to the use of the instrumental as the agent in passive constructions.

2.6. Some Consequences of the Use of Additional Substantives in Simple Sentences.

As the examples given above illustrate, case forms could be used independently of verbs, as well as to indicate their complements. The independent uses led to the development of adverbs from nouns which occurred commonly in one case form. The independent uses also led to the development of sentences with more than one nominal expression. In the course of time some of these uses came to be required, such as the use of the nominative to express the subject in most sentences. This use is mandatory in many contemporary languages, such as English; in contemporary English, grammars even require a subject in imperatives, such as wash, on the grounds that you must be posited at one stage of the derivation of such sentences in order to account for the reflexive in sentences like Wash yourself. In PIE, however, it is unnecessary to posit such a derivation, for the “reflexive” was indicated by a verbal suffix in the so-called middle, and was introduced through a Q category. Accordingly, nominative forms are not mandatory accompaniments of verbs, even of verbs in the imperative.

Like the nominative, the dative came to be mandatory with certain verbs, but only in certain uses. Next to the nominative it was the most widely used case in sentence patterns requiring the accusative, as with the root dō- ‘give’. In this way simple sentences resulted which were accompanied by three substantives:

68. RV 1.133.7. sunvānā́yéndro datāty ābhúvaṃ rayíṃ
to-the-presser-Indra he-gives ready wealth
‘To the soma-presser Indra gives ready treasure.’

Such sentences also resulted when other cases, such as the genitive, were used to supplement verbs:

69. RV 7.57.6. dádāta no amṛ́tasya prajā́yai
you-give our of-immortality descendants
‘Give immortality to our descendants.’

As is well known, the genitive in such uses had the meaning of a partitive object. The sentence also illustrates that verbs are mandatorily accompanied by at most three nominal expressions. This restriction gave rise to the classification of case categories into two sets: the grammatical and the local cases. The local cases consisted of the instrumental, the ablative, the locative, and possibly the dative; the grammatical consisted of the nominative, the accusative, and the dative, with the genitive functioning primarily as a form expressing relationship.

In addition to the sentence consisting of verb alone or of verb with from one to three mandatory cases in late PIE, substantives in the local cases could be employed with few restrictions. In the earliest texts such sentences are far less frequent than are sentences with only one or two substantival complements. But examples can be cited, like the following:

70. RV 9.1.7.
tám īṃ áṇvīḥ samaryá ā́
him him delicate for-the-festival Ptc.
gṛbhṇánti yóṣano dáśa
they-seize maidens ten
svásāraḥ pā́rye diví
sisters decisive on-day
  ‘Him the ten delicate maidens, the sisters [fingers], seize for the festival on the decisive day.’

In this sentence the verb is accompanied by a subject in the nominative, an object in the accusative, a dative, and a locative, to indicate the time of action.

71. RV 9.26.5. táṃ sā́nāv ádhi jāmáyo háriṃ hinvanty ádribhiḥ
it on-top onward sisters golden make-flow with-stones
  ‘[By crushing] it with stones the sisters make the golden juice flow on the surface.’

In this sentence the verb is accompanied by a subject and an object, plus an instrumental, ádribhiḥ, and a locative, sā́nāv (verse variant of sā́nāu) to indicate location. Other such examples could be cited from the early texts. They would indicate that the sentences of these texts rarely contained more than three nouns accompanying a given verb; when sentences are lengthy, the additional material generally consists of further elements coordinated with these or embedded.

2.7. Intonation Patterns of the Sentence.

As described in the previous sections, the sentence was characterized in PIE by patterns of order and by selection. Selection classes were determined in part by inflection, in part by lexical categories, most of which were covert. Some lexical categories were characterized 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. In addition to characterization by means of order and categories of selection, the sentence was also delimited by intonation based on variations in pitch.

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. The location of the high pitch is determined primarily from our evidence in Vedic; the theory that this was inherited from PIE received important corroboration from Karl Verner's demonstration of its maintenance into Germanic (1875). Thus the often cited correlation between the position of the accent in the Vedic perfect and the differing consonants in Germanic provided decisive evidence for reconstruction of the PIE pitch accent as well as for Verner's law, as in the perfect (preterite) forms of the root *deyk- ‘show.’

72.   PIE Vedic OE OHG
  1 sg. dedóika didéśa tāh zēh
  1 pl. dedikmé didiśimá tigon zigum

Studies to determine the details of accentuation in PIE and its relationship with patterns of accentuation in the dialects have led to many disagreements on details. But it is generally agreed that words were characterized on one syllable by a high pitch accent, unless they were enclitic, that is, unmarked for accent.

It is also agreed that accented words could lose their high pitch accent if they were placed at specific positions in sentences. Thus, 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.

Judging on the basis of loss of high pitch accent of verbs in them, independent clauses apparently were characterized by final dropping in pitch. For in unmarked order the verb stands finally in the clause.

73. Taittirīya Brāhmaṇa (TB) purodhā́m evá gacchati
priesthood verily he-attains
‘He attains the priesthood.’

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 Brāhmaṇa, 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 1970c:9):

74. TB gácchati pratiṣṭhā́m
‘He attains stability.’

Since gacchati in Example 73 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 1970c: 9):

75. TB yát stríyam upeyā́t || nírvīryas syāt #
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.

76. TB kásmai satrám āsmahe #
why indeed sacrifice we-perform  
‘Why should we perform sacrifice?’

77. TB stríyam úpeyāt #
‘He should not approach the woman.’

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:

78. Beowulf 2. þeodcyninga þrym gefrūnon
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.

79. Beowulf 6. egsode eorlas, syððan ǣrest wearð
he-terrified men since first he-was
‘He terrified men from the time he first was [found].’

80. Beowulf 30. þenden wordum wēold wine Scyldinga
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.

Among such patterns is the preference for enclitics in second position in the sentence (Wackernagel 1892:333). Words found in this position are particles, pronouns, and verbs, which have no accent in Vedic texts. This observation of Wackernagel's supports the conclusion that the intonation of the sentence was characterized 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 a line quoted above (Example 47):

81. RV 10.34.1. prāvepā́ bṛható 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 the next lines:

82. RV 10.34.1. pravātejā́ íriṇe várvṛtānāḥ
in-windy-place on-dice-board rolling
‘[born] in a windy place, rolling on the dice-board’

A simple sentence then consisted not only of a unit accompanied by an intonation pattern, but also of subunits or phrases. These were identified by their accent and also by patterns of permitted finals.

2.8. The Role of Modification as a Syntactic Marker.

As noted in the preceding section, single words or words accompanied by enclitics were identifiable on the basis of the incidence of accent; each such syntactic unit could have only one pitch accent. Such units were also identifiable by the patterns of inflectional syllables, as in:

83. ŚB tátaḥ saṃvatsaré púruṣaḥ sám abhavat prajā́patiḥ
from-it in-a year a-man together (he)-became he Prajapati
‘From it there developed in a year a man, Prajapati.’

As in this sentence, syntactic units like the subject, adverbial elements, and case forms were distinguished by accents and by final syllables which, as is well known, were weakened in articulation. This phenomenon is particularly prominent in Sanskrit (Whitney 1896:34-35). But from the restricted number of finals in other languages, such as only the vowels and the consonants n, r, s in Greek, we must assume that the final syllables of words had a special articulation already in PIE. This assumption requires the conclusion that words, or words accompanied by enclitics, were recognized as distinct syntactic units; in many instances they were phonologically distinguished by the articulation of the final syllable.

Details on the articulation of final syllables are a concern of morpho-phonemics. For syntactic purposes the essential concern is the fact that syntactic units smaller than the clause were identified in PIE. In general the obstruent series were reduced to one representative. In Sanskrit this was the voiceless member: p t ṭ k. In Greek however all of these were lost, and s was the only permitted final obstruent. In Latin, on the other hand, the permitted final dental was d, as in the Praeneste inscription:

84. Manios med fhe fhaced Numasioi
  Manios me reduplication he-made for-Numasios
  ‘Manius made me for Numasius.’

Subsequently other final consonants were permitted, including t in Classical Latin fēcit, upon the loss of final vowels. Furthermore, in Sanskrit, final s and final m were weakened, as illustrated in Example 70 with the standard transliteration for this weakened s. The phonologically based writing system for Sanskrit provides excellent evidence for these conclusions, as do the treatments of finals in other dialects.

Patterns of modulation then suggest that syntactic units which we may call phrases and also individual words were identified in the clause. Such syntactic units will be discussed further below.

2.9. Sentence-Delimiting Particles.

Indo-Europeanists and specialists in the various dialects have long noted the prominent use of particles in the early dialects (Delbrück 1888:23-24, 215-226, 472-546); but the discussions were primarily descriptive and etymological. Only with the discovery of Hittite did the syntactic use of particles for delimiting sentences become clear. Comparing the use in Celtic of particles to introduce some sentence types, Myles Dillon and Albrecht Götze proposed that the Hittite pattern was inherited from IE (Dillon 1947:15-24). It has subsequently become evident that reflexes of these patterns survived in other dialects as well, notably Sanskrit and Greek. Presumably particles had an important function in delimiting sentences. Yet their precise use needs much additional investigation. Among other matters the kinds of clauses that were introduced by particles must be determined. The use of particles in languages generally must also be investigated; although this is a general typological problem, it is important for IE studies, especially since particles have numerous syntactic uses. Moreover the use of particles changed considerably in the history of the dialects, and varied from style to style, as J. D. Denniston pointed out in his book The Greek Particles (1966a:lxiv-lxv, lxxv). The primary concern here is the use of particles to introduce clauses.

The particles concerned are PIE nu, so, to. Their homonymity with the adverb nu, nun and the anaphoric pronoun was one of the reasons earlier Indo-Europeanists failed to recognize them and their function. Yet Delbrück had already noted the clause-introducing function of Skt. sa (1888:215-216), as in:

85. ŚB
tásya tā́ni śīrṣā́ṇi prá cicheda. yát somapā́nam
his the heads off he-struck Ptc. what soma-drinking
ā́sa tátaḥ kapíñjalaḥ sám abhavat
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 1960:147-162); ta and šu occur primarily in the early texts, nu in the later, as illustrated in the following Old Hittite example (Otten and Souček 1969:38, § 22):

86. 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.’

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:

87. Odyssey 1.59.   oudé nu soí per
not-indeed and to-you but
entrépetai phílon ȇtor, Olúmpie  
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.

Many simple sentences in PIE would then be similar to those in Hittite and Vedic Sanskrit, such as those in the charming story taken by Delbrück from the Śatapathabrāhmaṇa (1878:63-64; see the excerpts in Chapter 3, Example 64). Among the simplest is the following:

88. ŚB tám índro didveṣa
and-him Indra he-hated
‘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 1960:63-64). The last sentence of the story shows the simple introductory particle:

89. ŚB sa tváṣṭhā cukrodha kuvín me putrám ávadhīd íti
Ptc. Tvaṣṭhar he-got-angry whether my son he-killed Ptc.
Tvaṣṭhar grew angry and said, “Has he killed my son?”’

Besides the use of sentence-delimiting particles, these examples illustrate the simplicity of PIE sentences. Of the fifteen sentences in the story, only two have more than one nominal form per verb, and these are adverbial as observed above, § 2.4.

90. ŚB eváṃrūpaṃ tenā́śanam ā́vayat
of-such-kind Ptc. he with-it-food he-enjoyed
‘For he enjoyed such kinds of food with it.’

Similar examples from the other early dialects could be cited, such as the Italic inscription of Praeneste (Example 84), or the Germanic Gallehus inscription:

91. Ek HlewagastiR HoltijaR horna tawido
  I Hlewagastir of-Holt the-horn I-made
  ‘I, Hlewagastir of Holt, made the horn.’

In these late texts, as noted above, § 2.4. and § 2.6, 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.