Wackernagel's Law and the Placement of the Copula Esse in Classical Latin

By J. N. Adams

This study, first published in 1994, considers the placement of the copula esse in Classical Latin and the possible relationship of its placement to Wackernagel's Law.

• Language: English
• Publisher: Cambridge Philological Society
• Release date: Aug 30, 2020
• ISBN: 9781913701246

Book preview

PREFACE

A version of this work was read at a meeting of the Cambridge Philological Society held at Christ’s College, Cambridge on 4 February 1993. I am grateful to those present on that occasion for their comments. I am particularly indebted to Professor P.H. Matthews, who read the whole manuscript and offered numerous very helpful observations and criticisms. A complementary paper, ‘Wackernagel’s law and the position of unstressed personal pronouns in Classical Latin’, is to appear in the Transactions of the Philological Society.

The aim of this work is to consider the placement of the copula esse in Classical Latin, and the possible relationship of its placement to Wackernagel’s law.

The material which must be considered in such a study is so extensive that I have deliberately restricted myself to prose of the Classical period, without attempting in any systematic way to identify possible diachronic developments in the placing of the copula. As will become clear, many questions have had to be left unanswered; I have from time to time raised issues which in my opinion deserve to be investigated further. Nevertheless the diverse patterns of placement of the copula in the Classical period can be reduced to a system of sorts.

1

Some types of enclitics in Latin

Latin had a number of postpositives which were possibly enclitics,¹ though the use of that term raises problems of definition. ‘Enclitic’ or more generally ‘clitic’ (which includes proclitics) tends to be used of an item which has no genuine independence. It links itself to a host which bears stress.² By ‘host’ I mean the word before the enclitic, on which it leans. The absence of an accent does not, however, appear to be a universal or necessary feature of enclitics.³ In a dead language the accentual pattern of host + posited clitic may in any case be a matter of guesswork, though there is a certain amount of evidence for the accentual behaviour of a few enclitics in Latin, as we shall see.

Wackernagel’s law is typically described in similar terms to those used by Watkins (1964, 1036): ‘One of the few generally accepted syntactic statements about IE is Wackernagel’s Law, that enclitics originally occupied the second position in the sentence.’⁴ The enclitics which are said to be placed second are a mixed lot: they include unstressed personal pronouns, usually in cases other than the nominative,⁵ sentence particles such as autem,⁶ various other particles such as ϰε in Greek,⁷ and sometimes the copula. Wackernagel himself (1892, 428–9) included the copula in his evidence from Latin.⁸ He quoted, for example, (1), where fore splits peraccommodatum and is thereby placed second in its clause, just as the pronoun mihi splits pergratum in (2):

(1)      Cic. Fam. 3.5.3 per fore accommodatum tibi

(2)      Cic. Att. 1.4.3 per mihi gratum est⁹

Other scholars who have implied that esse tends to be placed second because of the operation of Wackernagel’s law include Kroll (1925, 92, 95), Scherer (1975, 222), Wanner (1987, 84),¹⁰ Giacalone Ramat (1990, 176) and Pinkster (1990, 169).

Any definition (such as that of Watkins) which refers to the sentence or clause must now be modified in the light of Fraenkel’s work on colon division (1932, 1933, 1964, 1965, 1968). Fraenkel showed that long clauses can often be split into smaller units which he called ‘cola’.¹¹ Some enclitics, such as unstressed pronouns in Latin, may be placed second in any colon of a clause, and not exclusively in the first. For example in (3), if de triumpho autem is treated as a ‘short colon’, a sort of heading which might have been followed by a pause in speech, then me no longer violates a posited rule of second-position placement: it is in second place in the second colon, and it breaks up a syntactic unit to get there:

(3)      Cic. Att. 7.2.6 de triumpho autem // nulla me cupiditas umquam tenuit¹²

It is possible for a pronoun to be placed at the end of a sentence, yet still occupy the second position in its colon: e.g.

(4)      Cic. Att. 12.50 qua re cum poteris, // id est cum Sexti auctioni operam dederis, // reuises nos

(5)      Cic. Sest. 28 uenisse tempus // iis qui in timore fuissent // – coniuratos uidelicet dicebat // – ulciscendi se

There are various problems with conventional accounts of, or assumptions about, Wackernagel’s law which are now widely scattered in linguistic literature.

In the first place, the use of a single blanket term ‘enclitic’ obscures the fact that, even in a single language such as Latin, the items which tend to be labelled ‘Wackernagel enclitics’ are a mixed lot which may behave in rather different ways.

Secondly, once the assumption is made that enclitics mechanically occupy the second position in the clause or colon, then the first element or host tends to be treated as a matter of no importance. In this monograph I will on the contrary argue that the host is vitally important and that its position often determines the position of the enclitic.

Thirdly, it can be argued for Latin that different enclitics which more often than not come second in a colon, do so for different reasons.

Finally, the question must be asked whether all of the ‘Wackernagel enclitics’ are indeed regularly placed in second position in a colon in Latin, and whether that was in any case their original position. Evidence has recently been produced by Hale (1987) from Indo-Iranian languages suggesting that weak pronouns did not originally occupy the second position at all.¹³ A glance through any work of Cicero’s will turn up examples of unemphatic pronouns which, on a reasonable colon division, are later than second in their colon. What then determines their position? What is one to make of mihi in (6)?

(6)      Cic. Att. 11.5.4 neque illius // neque cuiusquam mihi praeterea // officium deest?

Pronoun position, however, is another story which I leave for another occasion (but see further below, (456)–(459).

For the purposes of this work I wish to distinguish between two, and perhaps three types of enclitics in Latin, without wishing to imply that these categories cover the field.

(i) Sentence enclitics. Sentence clitics ‘have scope over the entire sentence. They may mark the utterance as a question, as reported speech, as polite, as firmly believed or speculative, or … they may be connectives showing the relationship of the clause to what precedes or follows.’¹⁴ This definition, made in reference to a non-Indo-European language, could with a few modifications be applied to Latin. An obvious sentence enclitic in Classical Latin is autem. Another is -ne, which marks a clause as interrogative.¹⁵

Sentence enclitics in Latin are not as straightforward a category as may be supposed. Autem is perhaps alone in being regularly placed in second position at all periods. Enim and igitur, on the other hand, are often at the head of a sentence in early Latin.¹⁶ They seem to have shifted to the second position as a change of meaning took place.¹⁷ This fact suggests that it is unsatisfactory to set up an undifferentiated class of ‘Wackernagel sentence enclitics’ which supposedly gravitated mechanically towards the second position regardless of semantic considerations. The point has been well illustrated by Janson (1979), but disregarded by linguists keen to find linguistic universals at work. Another particle which moves between the first and second positions in response to semantic factors is ergo.¹⁸ There are in addition various particles which show an increasing tendency to take the second position in Classical Latin, without ever achieving fixed second-position placement (e.g. scilicet, uidelicet, nimirum, itaque, tarnen).¹⁹ It is not my task here to deal with sentence particles in detail. I would merely state that those listed above do not, on the face of it, look like a unified class. As a group they show no more than a tendency to adopt the second position.

(ii) ‘Internal’ or ‘non-sentential’ enclitics, recently called by Krisch (1990, 65) ‘Nicht-Wackernagel-Enklitika’. These act not on the sentence as a whole, but on their host. The host may occupy different positions in the sentence, and as a result the enclitic need not occupy the so-called ‘Wackernagel position’.²⁰ The clearest enclitic of this type in Latin is perhaps quidem.²¹ Its most distinctive use is contrastive,²² i.e. it is attached to one member of a contrasting pair: e.g.

(7)      Cic. Cat. 3.11 atque ille primo quidem negauit; post autem aliquanto, toto iam indicio exposito atque edito, surrexit

‘at FIRST he denied it, but LATER …’

(8)      Cic. Orat. 171 et apud Graecos quidem iam anni prope quadringenti sunt cum hoc probatur; nos nuper agnouimus …

‘among the GREEKS … WE … ’

Quidem may be in second position in the clause if its host is in first, but its placement is more varied than that of, say, enim. For quidem placed late in a clause in positions where a sentence enclitic such as enim could not possibly stand, note:

(9)      Cic. Brut. 18 non me hercule, inquit, tibi repromittere istuc quidem ausim

(10)    Cic. Fin. 4.43 itaque mihi uidentur omnes quidem illi errasse, qui …

There is some important evidence which suggests that, sometimes at least, quidem was unaccented. The combinations mequidem, tuquidem, siquidem (and a few other comparable cases) sometimes show shortening of the first vowel.²³ Latin monosyllables may have a long vowel or diphthong in final position (e.g. mē, tū, sī, dā, quī, quāē), but never a short vowel.²⁴ Short vowels in monosyllables must be followed by at least one consonant (dăt, ĕst etc.). It follows that the combination sĭ + quidem could not possibly have borne two accents (i.e. sĭ quĭdem) such that the two elements were treated as independent words (accentual units). Si would only retain a short vowel if it formed an accentual unity with quidem, i.e. sĭquĭdem (cf. such words as fắcĭlis); consequently quidem would have to be unaccented.

The host of quidem is generally the focus of the remark. Its character as such would be expressed in an English rendering by a falling pitch accent. I return to this notion of ‘focus’ later (5.1). Quidem can be described as a ‘focusing particle’. The contrastive or emphasising function of quidem has been well recognised, but in general it is true to say that the focusing role of some other enclitics in Latin (i.e. their role in marking the host as the focus of the clause) has not been systematically described. This essay will be mainly concerned with the relationship between enclitic and host, and it will be demonstrated that emphasising particles such as quidem are not the only focusing enclitics in Latin. Both the copula and the unstressed pronouns seem to have acquired such a function.

One final point can be made about quidem which will be occasionally relevant to our later discussion. Not infrequently it is displaced, so that it follows not its expected (focal) host, but another word in the clause. This phenomenon, which has been well documented,²⁵ does not take place haphazardly. Quidem often seems to be attracted to a pronoun, demonstrative or relative, apparently in defiance of the emphasis of the sentence: e.g.

(11)    Cic. Brut. 227 uerbis non ille quidem ornatis utebatur sed tamen non abiectis

Here (non) ornatis and (non) abiectis are in antithesis; one might have expected non ornatis quidem.

(12)    Cic. Fin. 1.16 cum miraretur ille quidem utrumque, Phaedrum autem etiam amaret

= miraretur quidem, or utrumque quidem (cf. (7) for quidem … autem)

Miraretur and amaret are contrasted, as are utrumque and Phaedrum.²⁶

(13)    Cic. Off. 1.33 decipere hoc