Pinkster, Harm (1942-) [1990], Latin Syntax and Semantics [info], xii, 320 p.: ill.; 24 cm. [word count] [Pinkster].
Previous Section

Next Section

7 Complex sentences (embedded predications on the sentence level) 7.1. Introduction 7.1.1 Working definition of the notion `complex sentence'

In the preceding chapters, examples have been given, without any further explanation, of subordinate clauses functioning as arguments or satellites within predications. In this chapter we will discuss such clauses in more detail. By way of introduction, I repeat a number of the examples already given above:

(1) Caesar cohortatur suos, ne animo deficiant (Caes. Civ. 2.43.1)

(2) me a portu praemisit, ut haec nuntiem uxori suae (Pl. Am. 195)

(3) (Romani) ex loco superiore … strage ac ruina fudere Gallos, ut nunquam postea nec pars nec universi temptaverint tale pugnae genus (Liv. 5.43.3)

(In (1) the italicized part of the sentence is an argument with the syntactic function Complement. In (2) the italicized part of the sentence is a satellite, with the function (Purpose) Adjunct. In (3) the italicized part of the sentence is a satellite as well, with the function (Result) Adjunct. Subordinate clauses with the function Disjunct may be found in crosssection 4.1.3–7 on pp. 34–7. Besides the predicate and its arguments, these clauses may in their turn contain satellites (e.g. numquam postea in (3)). In other words, a subordinate clause constitutes a predication which forms part of – is embedded in – the main (or matrix) predication.

Besides occurring as argument or satellite with their own function within a main predication, embedded predications may, as we have seen in chapter 6, also occur on the NP level. Examples are (4) and (5):

(4) cum … nuntium accepissem te mihi quaestorem obtigisse (`When I had received word that you had been assigned to me as quaestor', Cic. Fam. 2.19.1)

(5) hostes qui fugiunt non sunt timendi (`Enemies who flee need not be feared', see K.–St. II. 279)

-- 100 --

In (4) we find a so-called Accusative and Infinitive construction (AcI) which further specifies the content of nuntius (see p. 78). In (5) we see a relative clause in the function Attribute with hostes.

In this chapter we understand by `complex sentences' only those instances in which we find an embedded predication functioning as an argument or as a satellite with the main predicate. [1] Embedded predications may take the form of clauses introduced by a subordinator (e.g. ne, ut), but also that of an AcI, gerund, supine, infinitive, etc.

The grammars (Sz. XI; K.–St. II. 5) also treat as complex sentences cases like (6):

(6) ego hanc amo et haec me amat (`I love her and she loves me', Pl. As. 631)

In (6) the coordinator (so-called `coordinating conjunction') et connects two predications in one sentence. This instance differs from examples (1)–(3), however, in that here one predication is not embedded in the other as an argument or satellite. The grammars call (6) an instance of parataxis (`coordination', `Beiordnung') and (1)–(3) instances of hypotaxis (`subordination', `Unterordnung'). [2] To distinguish particles such as ne in (1) and ut in (2) and (3) from et, sed and other coordinators, they will be called `subordinators' (`subordinating conjunctions'). I will return to parataxis in chapter 12.

7.1.2 The modality of the embedded predication (semantic differences between types of embedded predication)

In chapter 10 on sentence types various types of independent sentence will be distinguished:

- declarative sentences (including `potential' sentences);

- interrogative sentences:

(a) questions with regard to the truthfulness of the statement (`yes–no' or `sentence' questions);

(b) questions introduced by a question word (`word' questions or `wh-questions');

- imperative sentences.

The difference between these sentence types often appears from their form, e.g. in Latin the presence of certain words and particles (num, -ne in questions) and in other languages the position of the finite verb in interrogative sentences; imperative or subjunctive verb forms in imperative sentences and subjunctive verb forms in potential and imperative sentences; the negation in imperative sentences.

Apart from these formal characteristics, some sentence types are subject to semantic restrictions, viz. (a) on the addibility of certain satellites, and (b) on the types of states of affairs that may occur in certain sentence types. For

-- 101 --

instance, certain Disjuncts cannot be added to imperative sentences and wishes (e.g. haud dubie, fortasse, see p. 33) or to interrogative sentences (cf. English: `Will he undoubtedly come?'). Incompatibility of sentence type and state of affairs plays a role in the case of non-controlled states of affairs. As a rule, such states of affairs do not occur in imperative sentences (see p. 17). The formal and semantic characteristics mentioned above are the expression of what is sometimes called the `modality' of each of these sentence types. In chapter 10 we return to this at greater length.

Semantic characteristics such as those mentioned above also hold for certain types of embedded predication. This appears from the fact that certain main predications do not allow embedding of certain predications. Thus, it is obvious that main predications with the Predicate to order can only embed predications with imperative modality. Main predications with the Predicate to tell, on the other hand, may embed both predications with declarative modality and predications with imperative modality, e.g.:

(7a) John told me (= said to me) that he would vacate the house

(7b) John told me (= ordered me) to vacate the house

The difference in modality between the embedded predications in (7a) (declarative) and (7b) (imperative) is reflected by the fact that the former allows the addition of a Disjunct such as perhaps, whereas the latter does not.

(7'a) John told me (= said to me) that he would perhaps vacate the house (e.g. if he finds a buyer)

(7'b) *John told me (= ordered me) perhaps to vacate the house

Like independent imperative predications, embedded imperative predications cannot be extended by the Disjunct perhaps ((7'b)), whereas declarative predications can ((7'a)). This also holds for Latin. In the treatment of embedded predications as arguments I return to the issue of possible restrictions on the modality of embedded predications.

7.2. Embedded predications functioning as arguments

In this section I give examples of embedded predications with one-place, two-place and three-place predicates, successively. A useful comparison may be made with the tables on pp. 21 ff., where the examples of arguments with the various predicates are for the most part noun phrases. It should be kept in mind, however, that on p. 21 ff. the syntactic construction (in terms of syntactic functions) is indicated, whereas the tables below show the formal expression of the embedded predications without indicating the syntactic structure. A number of the forms in which embedded predications occur will be further discussed in crosssection 7.4. Furthermore, we will see that the semantic and syntactic structure are not always parallel ( crosssection 7.5.) and that for a number of the

-- 102 --

predicates treated here as one-place we might also consider a treatment as two-place predicates. There are predicates with more than one meaning and with different frames. An example is necesse est (`it is necessary'). Most predicates can govern more than one form of embedded predication, in a number of cases with a difference in meaning. I return to this in crosssection 7.4. As appears from the tables, it is not true that certain types of embedded predication are reserved for one-, two- or three-place predicates. Finally, the selection of predicates dealt with here is highly limited. I do not pretend that the tables represent all possible combinations of types of embedded predication. Moreover, diachronic changes are not dealt with in a systematic way. [added 12-08: The tables are followed by examples. They may give an impression of the distribution of the specific constructions in authors and text types.]

Table 7.1 So-called impersonalia
(Main) PredicateForm of embedded predication
Embedded predicate: finiteEmbedded predicate: non-finite
(ut) non(ut) nequod/quia nonaquinbdependent questionInf.AcINcIDPtcGerSup
necesse este+-----+----
– : not attested in the period 200 BC–AD 100
+ : attested in the period 200 BC–AD 100
* : see note with predicate
(ut) : ut may sometimes be omitted
Inf. : infinitive (prolative infinitive)
AcI : `real' Accusative + Infinitivef
NcI : Nominative + Infinitive
DPtc : dominant participle construction (see crosssection 7.4.7)
Ger : any gerund (ive) construction
Sup : (second) supine (in -u)
a For (rare) quoniam = quod/quia see Sz. 628.
b Quin occurs only with negative main predications. See also on facere in crosssection 7.2.2.
c See K.–St. II. 239 for isolated accidit (ut) ne, and the like. We find the AcI especially in cases of accidit with an evaluation of the nature of the coincidence (e.g. inique or permirum); see K.–St. II. 240; Sz. 579). One might consider the possibility of calling inique etc. obligatory.
d For late embedded predications with [added 12-08: quod and] ut see TLL s.v. 535.80 ff.
e In the case of necesse est the so-called `inferential' use and the so-called `deontic' use must be distinguished. Example (11a) is an instance of deontic use, example (11b) an instance of inferential use (see Bolkestein 1980a: passim). Furthermore, there is also a two-place necesse est (mihi + inf.; see p. 108 example (24); see Bolkestein 1980a: 104). Necesse est is here treated as one indivisible idiom; etymological explanations of the form necesse are of little use with regard to the productive use of the expression.
f See for the distinction between `real' AcI and prolative infinitive crosssection 7.4.2.

-- 103 --

7.2.1 One-place predicates

The predicates discussed here are on the one hand the so-called `impersonalia' (refert, `it is of importance') and, on the other, expressions consisting of est or another copula and the neuter form of an adjective (utile est, `it is useful') or a noun (mos est, `it is customary'). [3] With these predicates the embedded predications function as Subject. After each table I occasionally discuss briefly those predicates which cannot with certainty be classified as one-place.

Examples of `impersonalia' (table 7.1)


(8a) tantum abest ut scribi contra nos nolimus ut id etiam maxime optemus (`We object so little to criticism against us that we even hope for it', Cic. Tusc. 2.4)

(8b) prorsus nihil abest quin sim miserrimus (`I am absolutely miserable', Cic. Att. 11.15.3)


(9a) capitis nostri saepe potest accidere ut causae versentur in iure (`It can often happen that matters which concern our position as a citizen are brought to court', Cic. de Orat. 1.181)

(9b) hoc loco percommode accidit quod non adest is qui … (`At this point it is a fortunate coincidence that he who … is not present', Cic. Caec. 77)

(9c) videte igitur quam inique accidat quia res indigna sit ideo turpem existimationem sequi … (`See how unjustly a bad reputation is the result of defending an unworthy cause', Cic. Caec. 8)


(10a) quo id factum nomine appellari oporteat constat (`It is clear how this act should be defined', Cic. Inv. 1.12)

(10b) quae si omnia e Ti. Coruncanii scientia … acta esse constarent (`If it were established that all this had been done in accordance with the knowledge of Ti. C.', Cic. Dom. 139) [4]

(10c) constat … ad salutem civium … inventas esse leges (`It is certain that the laws have been invented for the benefit of the citizens', Cic. Leg. 2.11)

necesse est

(11a) adhibeas necesse est omnem rationem (`You must use all your intellectual faculties', Q. Cic. Pet. 15)

-- 104 --

(11b) necesse est igitur legem haberi in rebus optimis (`So the law must needs be considered a very good thing', Cic. Leg. 2.12)


(12a) ut valeant refert (`It is important that they are healthy', Cato Agr. 73) [5]

(12b) ipsi animi magni refert quali in corpore locati sint (`Particularly for the soul it is of great importance in what kind of body they have been located', Cic. Tusc. 1.80)

(12c) neque enim refert videre quid dicendum sit (`For it does not matter very much if one sees what is to be said, but …', Cic. Brut. 110)

(12d) parvi refert abs te ipso ius dici aequabiliter (`It matters little whether you yourself judge impartially', Cic. Q. fr. 1.1.20)

The syntactic function of the embedded predication in an instance like (10b) is not easy to determine. Here an argument of the embedded predication is Subject of the main predication (quae): in this case the grammars speak of the `personal construction'. The Predicate of the embedded predication has been recorded in the figure as a prolative infinitive. In a case such as this, however, one could also regard constare as two-place, with one argument in the syntactic function Subject and one in the syntactic function Complement. I return to this phenomenon in crosssection 7.5.1. on p. 135. In other cases the embedded predication as a whole functions as Subject, as has already been said above (p. 103). Table 7.2 Copula + adjective
(Main) PredicateForm of embedded predication
Embedded predicate: finiteEmbedded predicate: non-finite
(ut) nonquod nonne nonquindependent questionInf.AcINcIDPtcGerSup
Abbreviations as for table 7.1.
a For (late) quod, quia, quoniam and ut see TLL s.v. 2115.32 ff.
b With utilis and other adjectives we do find gerund (ive) constructions, e.g. with ad, but I regard these as satellites.

-- 105 --

Examples of copula + adjective (table 7.2) [6]


(13a) difficile est … ut ad haec … studia … animus tam cito possit accedere (`It is difficult for the mind to occupy itself so rapidly with this (less serious kind of) literature', [7] Sen. Dial. 11.8.3)

(13b) non fuisse ei grave nec difficile eam causam excipere (`That it had been neither hard nor difficult to accept an exception in this case', Cic. Inv. 2.130)

(13c) difficile est in Asia … ita versari nostrum imperatorem, ut … (`It is difficult for a Roman commander to behave in Asia in such a way that …', Cic. Man. 64)

(13d) ut in comparando difficile ad eligendum sit quid maxime velis (`So that it is comparatively difficult to choose what you want most', Cic. Rep. 1.55)

(13e) difficile est hoc genus exornationis inventu (`This kind of figure of speech is difficult to find', Rhet. Her. 4.39)


(14a) non potest esse dubium quin id sit summum … bonorum omnium (`There can be no doubt that this (viz. pleasure) is the highest good', Cic. Fin. 1.54)

(14b) primo nobis fuit dubium quid ageremus (`At first we were in doubt as to what to do', Cic. Ver. 4. 138) [8]

(14c) perisse me una haud dubiumst (`Doubtless I, too, am finished', Ter. Hec. 326)


(15a) quid tam inusitatum quam ut … eques Romanus ad bellum maximum … mitteretur (`What is as uncommon as to send a Roman knight into a most serious war', Cic. Man. 62)

(15b) esta ita inusitatum regem reum capitis esse ut … (`It is so uncommon for a king to stand trial for a capital offence, that …', Cic. Deiot. 1)


(16a) si verum est – quod nemo dubitat – ut populus Romanus omnes gentes virtute superarit (`If it is true – as is doubted by no one – that the Roman people have been superior to all nations in courage', Nep. Han. 1.1)

(16b) nec verum est … idcirco initam esse cum hominibus communitatem (`It is not true that for this reason relationships have formed among human beings', Cic. Off. 1.158)

-- 106 --


(17a) numquam igitur est utile peccare (`So it is never useful to sin', Cic. Off. 3.64)

(17b) Miloni etiam utile fuisse Clodium vivere (`That it was even in M.'s interest that C. was alive', Cic. Mil. 52)

Many predicates consisting of copula + adjective occur with a nominal constituent in the dative. I regard such constituents as satellites with the semantic function Beneficiary (e.g. with difficilis and utilis).

Apart from (13e), I have only given examples of so-called `impersonal' constructions of esse + adjective. I regard (13e) and similar examples as equivalent to the other instances, although here, too, we should formally speak of a so-called personal construction. Here, therefore, mutatis mutandis the same applies as was remarked on p. 104 with respect to constarent in example (10b). Looking only at the syntactic structure, one might also speak of two-place difficilis. (13e) is thus described in a different way from cupidus in (25a–b) and nescius in, for example:

(18) iratum te regi Deiotaro fuisse non erant nescii (`They knew very well that you had been angry with King D.', Cic. Deiot. 8)

Nescius is two-place; see further crosssection 7.5.2. Table 7.3 Copula + noun
(Main) predicateForm of embedded predication
Embedded predicate: finiteEmbedded predicate: non-finite
(ut) non(ut) nene nonquod/quia nonquindependent questionInf.AcINcIDPtcGerSup
Abbreviations as for table 7.1

Examples of copula + noun (table 7.3)


(19a) facinus est vincire civem Romanum (`It is a crime to put a Roman citizen in irons', Cic. Ver. 5.170)

-- 107 --

(19b) occisus dictator … pulcherrimum facinus videretur (`The murder of the dictator seemed a most beautiful crime', Tac. Ann. 1.8.6)


(20a) cum … mos esset Graecis … ut ei qui vicissent tropaeum … statuerent (`Since it was a habit among the Greeks that victors erected a monument', Cic. Inv. 2.69)

(20b) Magorum mos est non humare corpora suorum, nisi … (`It is a habit among the Magi not to bury their dead unless …', Cic. Tusc. 1.108)

(20c) qua (oratione) mos est Athenis laudari in contione eos qui sint in proeliis interfecti (`It is a custom in Athens to honour with this speech those who have been killed in battle', Cic. Orat. 151)

(20d) (fuit) mos vero liberos … obiciendi saevissimis earum (`They did in fact expose their children to the most frightful among them (viz. snakes)', Plin. Nat. 7.14)


(21a) sed iustitiae primum munus est, ut ne cui quis noceat (`The first duty of justice is to make sure that no one inflicts damage on someone else', Cic. Off. 1.20)

(21b) munus autem animi est ratione bene uti (`It is the duty of the soul to make good use of reason', Cic. Tusc. 3.15)


(22a) videtur tempus esse ut eamus ad forum (`It seems to be time to go to the forum', Pl. Mil. 72)

(22b) tempus esset iam de ordine argumentorum … aliquid dicere (`Then it would be time to say something about the order of the arguments', Cic. de Orat. 2.181)

(22c) tempus est … iam hince abire, me ut moriar, vos ut vitam agatis (`It is already time to leave, for me in order to die, for you in order to live your lives', Cic. Tusc. 1.99)

(22d) navigandi nobis … tempus esse (`That it is time for us to sail', Cic. Ac. 2.147)

(22e) si tempus est ullum iure hominis necandi (`If there is any one occasion for justifiable murder', Cic. Mil. 9) [9]


(23) alterum est vitium quod quidam nimis magnum studium … conferunt (`The second error is the fact that some devote too much attention to …', Cic. Off. 1.19) [10]

-- 108 --

A survey of the constructions of copula + noun may be found in K.–St. (II.272–3). Incidentally, here, too, we are confronted by the problem that it is difficult to distinguish between esse as a copula and existential esse. In table 7.3 it has not been attempted to make a distinction along these lines.

7.2.2 Two-place (main) predicates

In this section we deal with the same constructions as in the preceding one. I give one example of the so-called impersonal constructions (necesse est, propositum est + dative + infinitive) and of the adjectives. Adjectives which also have an impersonal construction have already been discussed in crosssection 7.2.1. (e.g. difficile). [11] Table 7.4
(Main) predicateForm of embedded predication
Embedded predicate: finiteEmbedded predicate: non-finite
(ut) non(ut) nene nonquod/quia nonquinotherdependent questionInf.AcINcIAcPaDPtcGerSup
necesse estb+------+------
Abbreviations as for table 7.1
a AcP: accusative + present participle.
b See table 7.1, note e.
c I am referring to the construction audire aliquem cum as in ex. (26a).
d TLL s.v. 349.42 ff. mentions late instances of AcI (Lucifer, Chiron); in Augustine in psalm 36, serm. 2.1 the AcI is dependent on optat.
e non metuo quin: see K.-St. II.256, A.4. AcI constructions only late.
f Instances with quin i.a. Cic. Rep. 1.50 quin serviant … fieri non potest (`It is impossible that they will not be slaves') (TLL s.v. 106.77 ff.). For quo and quominus TLL s.v. 106.58 ff.; see also TLL s.v. 107.15 ff.
g In Plin. Ep. 9.33.2 the AcI is governed by statui (contra TLL, s.v. deliberare 440.65 ff.).

Examples of constructions with two-place (main) predicates (table 7.4)


(24a) pro hoc mihi patronus sim necesse est (`For him I must be a `patronus", Pl. Poen. 1244)

-- 109 --

(24b) non est omnibus stantibus necesse dicere (`It is not necessary that all should stand up to speak', Cic. Marc. 33)


(25a) mortem … timens cupidusque moriri (`Fearing death and longing to die', Ov. Met. 14.215)

(25b) tu qui valde spectandi cupidus esses (`You who were very much eager to see them', Cic. de Orat. 1.162)


(26a) saepe enim soleo audire Roscium, cum … dicat … (`Often I hear R., when he says …', Cic. de Orat. 1.129)

(26b) nemo fere vestrum est quin quemadmodum captae sint a M. Marcello Syracusae saepe audierit (`Hardly anyone among you has not often heard how Syracuse was captured by M.', Cic. Ver. 4.115)

(26c) M. vero Scaurus, quem non longe ruri apud se esse audio (`M.S., who, as I am told, is not far from here at his country estate', Cic. de Orat. 1.214)

(26d) bellum … ante audierunt geri quam parari (`They heard that the war was being waged before they heard that it was being prepared', Cic. Lig. 3)

(26e) Bibulus nondum audiebatur esse in Syria (`There was no word so far of B. being in Syria', Cic. Att. 5.18.1)

(26f) idque Socratem … audio dicentem … (`I hear Socrates say that …', [12] Cic. Fin. 2.90)


(27a) Helvetii … si perrumpere possent conati … telis repulsi hoc conatu destiterunt (`The Helvetians tried to force a break-through; pushed back by spears they desisted from their attempt', Caes. Gal. 1.8.4)

(27b) quod sibi probare non possit, id persuadere alteri conetur (`He is trying to persuade the other of something he cannot prove himself', Cic. Q. Rosc. 4)


(28a) tibi cum omnia mea commendatissima esse cupio tum nihil magis, quam ne tempus nobis provinciae prorogetur (`I wish all my affairs to have your attention, and above all that my term as provincial administrator is not extended', Cic. Fam. 2.8.3)

(28b) tu vellem ego vel cuperem adesses (`I would want, or rather, wish you to be here', Cic. Att. 2.18.4)

-- 110 --

(28c) qua exposita scire cupio quae causa sit cur Zeno … (`After this exposition I wish to know what the reason was why Zeno …', Cic. Fin. 4.19) [13]

(28d) liberos suos … beatos esse cupiat (`He wishes his children to be happy', Cic. Inv. 1.48) [14]

(28e) ego me cupio non mendacem putari (`I do not wish to be regarded as a liar', Cic. Leg. 1.4)


(29a) tu malim … actum ne agas (`I would prefer you not to act upon matters that have been closed', Cic. Att. 9.18.3)

(29b) malo non roges (`I had rather you did not ask questions', Cic. Tusc. 1.17) (contrastive non)

(29c) virtutum in alia alius mavult excellere (`Everyone wants to excel in another virtue', Cic. Off. 1.115)

(29d) sed fortasse maluit … omnium … esse princeps (`But perhaps he wanted to be the best of all', Cic. Brut. 151)

(29e) Afer aut Sardus …, si ita se isti malunt nominari (`African or Sardinian, if that is what they prefer to be called', Cic. Scaur. 15)


(30a) illud gaudeo, quod … aequalitas vestra et … abest ab obtrectatione <et> invidia … (`I am glad that your equality does not cause disparaging behaviour or jealousy', Cic. Brut. 156)

(30b) et quom te gravidam et quom te pulchre plenam aspicio, gaudeo (`And when I see you pregnant and beautifully full, I rejoice', Pl. Am. 681)

(30c) abs quivis homine … beneficium accipere gaudeas (`You can look forward to receiving favours from anyone, no matter who', [15] Ter. Ad. 254)

(30d) venire tu me gaudes (`You are happy about my arrival', Pl. Bac. 185)

(30e) cum … valde absoluto Scaevola gauderet (`When he was very pleased with the acquittal of S.', Cic. de Orat. 2.281)


(31a) atque etiam id ipsum, quod tu scribis, metuebam, ne a me distrahi non posses (`And I was afraid of the very thing you mentioned, viz. that you could not tear yourself away from me', Cic. 1.3.4)

(31b) ego ... non tam veteranos metuendos nobis arbitror quam quid ... cuncta Italia ... sentiat (`I think that we should not fear so much the veterans as what the whole of Italy feels', Cic. Phil. 11.39) [added 12-08]

(31c) ego nec tumultum nec mori per vim metuam tenente Caesare terras (`I will fear neither unrest nor violent death as long as Caesar governs the world', Hor. C. 3.14.14–16)

-- 111 --

(31d) an metuit (anima) conclusa manere in corpore putri et … (`Or is the soul afraid to be left behind, in a rotting body and …', Lucr. 3.773)


(32a) nam quod rogas curem ut scias quid Pompeius agat (`For as to the fact that you ask me to let you know how P. is', Cic. Att. 7.12.1)

(32b) is curavit quod argumentum ex Dionysio ipse sumpsisset ex eo ceteri sumerent (`He made sure that the argument which he had taken from D. was taken from him by the others', Cic. Ac. 2.71)

(32c) ea nolui scribere quae nec … nec docti legere curarent (`I did not want to write those things … which learned men do not find worth reading', Cic. Ac. 1.4)

(32d) nec vera virtus … curat reponi deterioribus (`True courge does not wish to return into people who are less good', Hor. C. 3.5.29–30)

(32e) ut natura et procreari vellet et diligi procreatos non curaret (`That nature both wanted procreation and did not make sure that those who are born receive love', Cic. Fin. 3.62)

(32f) … eum hominem occidendum curavit (`He had this man killed', Cic. S. Rosc. 103) [16]


(33a) splendor vester facit ut peccare sine summo rei publicae detrimento … non possitis (`Your splendid record makes it impossible for you to err without doing serious harm to the state', Cic. Ver. 1.22)

(33b) fecisti ut ne cui innocenti maeror tuus calamitatem … adferret (`You have ensured that your grief did not cause a disaster for an innocent person', Cic. Clu. 168)

(33c) bene facis … quod me adiuvas (`You do well to help me', Cic. Fin. 3.16)

(33d) visum est faciendum … vos certiores facere (`It seems necessary to inform you', Sulp. in Cic. Fam. 4.12.1) [17]

(33e) (actio) tales … oratores videri facit, quales ipsi se videri volunt (`The delivery causes the orators to make the impression they want to make', Cic. Brut. 142) [18]


(34a) brevior iam in scribendo incipio fieri (`I am already beginning to write more concisely', Cic. Att. 5.6.2)

(34b) nunc quoniam de re publica consuli coepti sumus (`Now that we are again asked for political advice', Cic. Div. 2.7)

-- 112 --


(35a) deliberant quid agant (`They are deliberating as to what to do', Rhet. Her. 3.5.8)

(35b) cum deliberassent nobiscum bellum gerere (`While they had considered waging war against us', Rhet. Her. 4.13)

(35c) non de absolvendo Apronio deliberarent (`They would not have considered the release of A.', Cic. Ver. 3.31)

A number of nouns (such as dies (`day'), comitia (`elections'), locus (`place', `opportunity')) may be construed with esse and a gerund(ive) construction in the dative:

(36) dies is erat legitimus comitiis habendis (`According to law, this was the day for holding the elections', Cic. Ver. 2.129, cf. K.–St. I.748–9)

The same nouns can also be used with transitive verbs, again with a gerund(ive) construction in the dative:

(37) quibus (imaginibus) non … imitandis … tibi locum ullum reliquisti (`You have left yourself no room for imitating these', Cic. de Orat. 2.226)

Such expressions have been left out of account here, as well as opus est + dative.

7.2.3 Three-place predicates Table 7.5
(Main) predicateForm of embedded predication
Embedded predicate: finiteEmbedded predicate: non-finite
(ut) non(ut) nene nonquod/quia nonquinotherdependent questionInf.AcINcIAcPDPtcGerSup
Abbreviations as for table 7.4.
a See Petersmann (1977: 215-6) about alternative constructions for the AcI; TLL s.v. dicere 985.78 ff. See also crosssection 7.4.4 (p. 129) on differences in meaning between the various constructions with dicere.
b From Tertullian onwards.
c See note 11.
d See note 11.
e Reflexive (sibi persuasum est) has been left out of account. The Oxford Latin Dictionary mentions one instance with quod:

- persuaserunt Publio Maevio quod hereditas ad eum pertineret (`They convinced P.M. that the inheritance pertained to him', Scaev. Dig. 13.5.31).

f For late quod see TLL s.v. 766. 16 ff.

-- 113 --

Examples of constructions with three-place predicates (table 7.5)


(38a) dicebam, pater, tibi, ne matri consuleres male (`I told you, father, to take good care of mother', Pl. As. 938)

(38b) dices (eis) … paulum proferant auctionem (`Tell them to put off the auction for some time', Cic. Att. 13.12.4)

(38c) dixi quia mustella comedit (`I said that a weasel had eaten them', Petr. 46.4)

(38d) non … solum Torquatus dixit, quid sentiret, sed etiam cur (`T. did not only say what he thought, but also why', Cic. Fin. 2.3)

(38e) (Augustus patribus) sedentibus valere dicebat (`A. said farewell to the senators while they remained seated', Suet. Aug. 53.3)

(38f) quas (minas) hodie adulescens Diabolus ipsi daturus dixit (`Which the young D. has said he will give her today', Pl. As. 634)

(38g) omnes in iis sedibus quae erant sub platano consedisse dicebat (`He said that all had sat down in the chairs under the plane-tree', Cic. de Orat. 1.29)

(38h) dici mihi memini … L. Crassum … se … contulisse (`I recall that L. C. was said to have gone to …', Cic. de Orat. 1.24)

(38i) si dici possit ex hostibus equus esse captus (`If a horse can be said to have been captured from the enemy', Cic. Inv. 1.85)


(39a) eum (lenonem) ego docebo …, ut sibi esse datum argentum dicat (`I will make clear to him that he must say that the money was given to him', Pl. Epid. 364–5)

(39b) vos docebit, qualis sit L. Flaccus? (`Will he tell you what kind of man L. F. is?', Cic. Flac. 8)

(39c) neque … conamur docere eum dicere, qui loqui nesciat (`We do not try to teach someone who cannot talk how to speak', Cic. de Orat. 3.38)

(39d) docuerunt … me periti homines … transferri nomen … non potuisse (`Experts have told me that guilt cannot be transferred', Cic. Fam. 5.20.3)

(39e) ut … minime … pecuniae cupidus fuisse doceatur (`That it is argued, that he had by no means desired money', Cic. Inv. 2.36)


(40a) impedior … dolore … ne … plura dicam (`I am prevented by sorrow from saying any more', Cic. Sul. 92)

-- 114 --

(40b) … ne quid impediare quin ad hanc utilitatem pariter nobiscum progredi possis (`In order that you are not hindered in acquiring this skill together with me', Rhet. Her. 3.1)

(40c) cur, quo setius omnia scribant, impediuntur modestia? (`Why does modesty keep them from writing down everything?', Rhet. Her. 4.4)

(40d) me … impedit pudor ab homine … haec … exquirere (`Embarrassment keeps me from asking someone this', Cic. de Orat. 1.163)

(40e) quas … ad capiendam fugam … infirmitas impediret (`Whom weakness prevented from fleeing', Caes. Gal. 7.26.3)


(41a) ei palam imperat ut omne argentum … conquirendum curaret et ad se adferendum (`He openly orders him to have all the silver brought together and brought to himself', Cic. Ver. 4.50)

(41b) imperat (Labieno) … eruptione pugnet (`He orders L. to make a sortie', Caes. Gal. 7.86.2)

(41c) non imperabat (obstetrix) coram, quid opus facto esset puerperae (`She did not herself give orders as to what was to be done for the mother', Ter. An. 490)

(41d) imperavi egomet mihi omnia adsentari (`I have ordered myself to agree to everything', Ter. Eu. 252–3)

(41e) Cleomenes … vela fieri, praecidi ancoras imperavit (`C. gave the order to raise the sails and to cut the anchors', Cic. Ver. 5.88) [19]

(41f) in has lautumias … deduci imperantur (`People are ordered to be brought to these quarries', Cic. Ver. 5.68) [20]


(42a) Telebois iubet (Amphitruo) sententiam ut dicant suam (`A. orders the T. to tell him their opinion', Pl. Am. 205)

(42b) curriculo iube in urbem veniat (`Order him to come to town with great haste', Pl. Mos. 930)

(42c) te lex Terentia … frumentum emere … iussit (`The lex Terentia ordered you to buy corn', Cic. Ver. 3.173)

(42d) (P. Scipio) iubet omnia conquiri (`P.S. ordered a general search to be made', Cic. Ver. 4.73)

(42e) cum alterius populi maiestas conservari iubetur (`When it is ordered to preserve the greatness of a nation', Cic. Balb. 36)

-- 115 --


(43a) patri persuasi, ut aes alienum filii dissolveret (`I persuaded the father to annul his son's debt&rsquo, Cic. Phil. 2.46)

(43b) huic Sp. Albinus … persuadet … regnum Numidiae ab senatu petat (`Sp. A. persuaded him to ask the senate for control over Numidia', Sal. Jug. 35.2)

(43c) quibus persuasum est foedissimum hostem iustissimo bello persequi (`Who are planning to pursue a most horrible enemy in a completely justifiable war', Cic. Phil. 13.35) [21]

(43d) hos homines tu persuadebis ad honorem … tuum pecunias maximas … contulisse (`You are trying to make us believe that these men have brought together enormous amounts of money in your honour', Cic. Ver. 2.157)


(44a) illud me praeclare admones …, ne nimis indulgenter et ut cum gravitate potius loquar (`You rightly admonish me to do this, viz. not to speak too leniently, but rather with some dignity', Cic. Att. 9.9.2)

(44b) maxime autem admonendus, quantus sit furor amoris (`But above all he must be warned of how great the madness of love is', Cic. Tusc. 4.75)

(44c) nonne te … Q. illa Claudia aemulam domesticae laudis … esse admonebat? (`Did not even the notorious Q. C. incite you to be her rival for "household praise" ', Cic. Cael. 34)

(44d) tantum te admonebo … praesentibus te his daturum (salutem) (`I merely remind you of the fact that you will bring salvation to all present here', Cic. Lig. 38)

(44e) admonitus sum ab illo … illis de rebus dici posse (`It has been pointed out to me by him that these things can be discussed', Cic. Q. fr. 3.5.1)

(44f) nostrique detrimento admonentur diligentius … stationes disponere (`Our soldiers were incited by the disaster to post sentries with more care', Hirt. Gal. 8.12.7)

In a number of the instances cited here no third argument is presupposed ((38c–d); (41e); (42d)). For the sake of convenience I have presented all examples of each verb here (see also n. 11).

7.3. Embedded predications functioning as satellites

This section deals with embedded predications functioning as satellites.

-- 116 --

Among these are what the grammars call `adverbial subordinate clauses'. A number of these are formally identical with the constructions we have seen in crosssection 7.2. This formal similarity is the reason why many scholars assumed and, in fact, still assume, a historical relationship between the occurrence of embedded predications as arguments and as satellites (see crosssection 7.6.). In crosssection 7.3.1 I discuss satellites in the function Adjunct, in crosssection 7.3.2 satellites in the function Disjunct. In a final section ( crosssection 7.3.3) I address the question of how subordinate clauses such as that introduced by ubi in (45) are to be described:

(45) exsilium ibi esse putat, ubi virtuti non sit locus (`He thinks that exile is there, where there is no room for virtue', Cic. Mil. 101)

7.3.1 Satellites in the function Adjunct

In this section, as in crosssection 3.3. (pp. 28 ff.), I treat the satellites according to the semantic functions they fulfil. The enumeration of semantic functions does not pretend to be exhaustive.

(i) Manner

Manner clauses introduced by ut, quemadmodum (`just as') are left out of account. K.–St. (I. 752) discuss the use of gerund and gerundive as Manner Adjunct. An example is (46):

(46) qui partis honoribus eosdem in foro gessi labores quos petendis (`I, who have made the same efforts on the forum after having held all offices as when I was striving for those offices', Cic. Phil. 6.17)

It is difficult to analyse instances of this kind: in this particular case I would be more inclined to speak of a Time Position Adjunct. [22]

(ii) Instrument

(47) pultando … confregi fores (`I have forced the doors by battering at them', Pl. Mos. 456)

(48) quae ipse in meis fundis colendo animadverti (`Which I have noticed myself on my estates by cultivating them', Var. R. 1.1.11)

K.–St. mention (48), together with three other cases, as an example of the `causal' use of the gerund(ive). I, however, treat (48) as equivalent to (47), as an Instrument Adjunct. The difference between the two instances is that (47) in all likelihood involves intention, whereas (48) does not. See also (viii) below. [23]

(iii) Degree

Subordinate clauses expressing Degree are as a rule introduced by quantopere, quam and the like (`how much').

-- 117 --

(iv) Position in time

Apart from clauses introduced by subordinators such as postquam (`after'), cum (`when'), quotiens (`whenever'), of which no examples are given here, we also find the gerund(ive) construction:

(49) quod … in redeundo … auspicari esset oblitus (`Because on his return he had forgotten to study the omens', Cic. N. D. 2.11)

For examples see K.–St. I. 753.

(v) Duration

Clauses with this function are as a rule introduced by subordinators such as quamdiu and dum (`as long as').

(vi) Position in space

For clauses introduced by ubi (`where'), unde (`whence'), qua (`along what route'), quo (`whereto') and the like, see crosssection 7.3.3.

(vii) Accompanying circumstances

(50) Bellum Gallicum … C. Caesare imperatore gestum est (`A war was fought in Gaul under the command of C. C.', Cic. Prov. 32)

(51) qui tranquillo mari gubernare se negent posse (`Who say they cannot sail on a calm sea', Cic. Rep. 1.11)

(52) (Cethegus) recitatis litteris … repente conticuit (`After I had read the letter aloud, C. was suddenly silent', Cic. Catil. 3.10)

(53) concursus est ad Templum Concordiae factus senatum illuc vocante Metello consule (`There was a run on the temple of Concordia when the consul M. convened the senate there', Cic. Dom. 11)

In chapter 3 (p. 29) no definition is given of the semantic function `Accompanying Circumstances'. As for the instances given above, one might say that in comparison with the other semantic functions the function `Accompanying Circumstances' is a kind of residual function, which indicates less explicitly than the other functions that the main predication was realized in the light of what is expressed in the embedded predication.

In the four instances cited here both the states of affairs and the entities (persons) involved in main and embedded predication are different, just as in the example on p. 29 assensu omnium refers to an action performed by persons not involved in the main predication. (N. B. nevertheless identity of the Agent is often presupposed: in urbe capta hostes abeunt (`after taking the city the enemies go away') the hostes are of course meant to be understood as those who have taken the city.) I assume that the ablative absolute construction, with which we are dealing here, is in reality nothing else than a Dominant

-- 118 --

participle construction functioning as a satellite with regard to the remainder of the predication (see crosssection 7.4.7. on p. 132 ff.).

(viii) Motive

In the grammars the notions `Cause', `Motive' and `Purpose' are often used without being clearly distinguished. Up to now I have used `Cause' for the non-human, most often inanimate entity responsible for a certain state of affairs. In traditional grammars also `Cause' is used as a label for this type of entity. I use the term `Motive' (German: `Graund') for the considerations which induce someone to do something, where, again, Cause is being used by other linguists. The difference between Cause and Motive can be made clear with the aid of the English paraphrases `due to' and `because of'. I reserve the term `Purpose' for that which someone intends his actions to accomplish. In some cases the reason why someone does something is at the same time that which he wants to accomplish. The main difference seems to be that `Purpose' refers to possible, future, situations, and `Motive' to real, past or present situations (see below). In specific contexts it may be difficult to decide whether one is dealing with Purpose or Motive.

The most common form of Motive Adjuncts is a clause introduced by one of the subordinators quod, quia, quoniam, cum. As correlating adverbs we find propterea, idcirco, ideo. [24] An example of a subordinate clause introduced by a subordinator (quoniam) is (54):

(54) sane gaudeo quod te interpellavi quoniam quidem tam praeclarum mihi dedisti iudicii tui testimonium (`I am very glad that I have interrupted your speech, because you have given me such splended proof of your appreciation', Cic. Leg. 3.1)

Embedded predications in function Motive Adjunct are also found in the form of preposition phrases (ob, propter, de (`on account of') + Dominant participle). An example of a Dominant participle construction introduced by a preposition is (55). Further examples may be found in K.–St. (I.767).

(55) qui ob eam (amicitiam) summa fide, constantia iustitiaque servatam maximam gloriam ceperit (`Who by the very faithful, constant and just preservation of friendship has achieved the greatest fame', Cic. Amic. 25)

We also find constructions of ob, etc. + gerund(ive) construction as Purpose satellites. And example is (56):

(56) existumans Iugurtham ob suos tutandos in manus venturum (`Thinking that I. would open combat in order to protect his subjects', Sal. Jug. 89.2)

Unlike the embedded predication in (55), the embedded predication in (56) does not refer to a state of affairs which has actually obtained (which, in other words, is factive, see p. 79), but to a state of affairs which is intended and will

-- 119 --

possibly obtain. Ob, propter, etc., therefore, occur with two constructions: the `factive' Dominant participle (Motive) and the `non-factive' gerund(ive) (Purpose). In (ix) on Purpose we will see that causa and gratia occur almost exclusively with the gerund(ive).

Just now it was remarked that Purpose and Motive seem to differ in that Purpose refers to states of affairs which do not yet obtain and of which it is not certain that they will obtain. Besides this difference, there are also similarities. Just now, we saw that a number of prepositions is used for both types of Adjunct. Both types of Adjunct may occur as the answer to the question cur? (`Why?'). Moreover, the two types of Adjunct occur in coordination (e.g. Cic. Att. 3.4). [25]

(ix) Purpose

As Purpose Adjuncts we find several constructions. The most common constructions are:

(a) ad + gerund(ive) construction (see K.–St. I.749 ff.); [26]

(b) causa/gratia (`on account of') + gerund(ive);

(c) subordinate clauses introduced by ut (`in order that'; negation ne), quo, qui (see K.–St. II. 232 ff.). Correlating expressions are e.g. idcirco, ideo (`therefore'), eo consilio (`with this intention') and ea condicione (`on this condition'), etc;

(d) supine in -um (see K.–St. I.721 ff.).

(57) legum … idcirco … servi sumus, ut liberi esse possimus (`We are slaves of the law, in order that we can be free', Cic. Clu. 146)

(58) subacto mihi ingenio opus est, ut agro non semel arato, sed novato et iterato, quo meliores fetus possit et grandiores edere (`I need a somewhat developed talent, like a field that has not been ploughed only once, but several times, in order that it may produce better and more copious crops', Cic. de Orat. 2.131)

(59) cum ad vim faciendam quidam armati venissent (`When certain men had come armed in order to commit acts of violence', Cic. Inv. 2.59)

(60) etsi admonitum venimus te, non flagitatum (`Though we come to give you a reminder, not to insist', Cic. de Orat. 3.17)

Besides these more or less common constructions we also find:

(e) infinitive, particularly in poetry and prose from Livy onwards, with predicates of a number of semantic classes (see K.–St. I.680 ff.)

(f) the dative of the gerund(ive) (see K.–St. I.749).

Examples of these constructions are:

-- 120 --

(61) illa … abiit aedem visere Minervae (`She left in order to visit the temple of Minerva', Pl. Bac. 900–1

(62) (Ganymedes) Iovi bibere ministraret (`G. served I. something to drink', Cic. Tusc. 1.65) [27]

(63) serviendae servituti ego servos instruxi mihi, hospes, non qui mi imperarent (`I have trained my slaves to serve me, my friend, not to give me orders', Pl. Mil. 745–6)

Further examples of a pre- and postposition phrase are (64) and (65), respectively:

(64) ad dicendum si quis acuat … iuventutem (`If someone were to train young people to become orators', Cic. Orat. 142)

(65) (filiam) necavere, seu ut vi capta potius arx videretur, seu prodendi exempli causa ne … (`They killed the daughter, either in order to make the citadel seem to have been taken by force, or to set an example', Liv. 1.11.7)

As was already remarked under (viii) above, causa and gratia frequently occur with the gerund(ive) construction but not with the Dominant participle construction, and may, therefore, be regarded as unequivocal markers of embedded predications with the semantic function Purpose. [28]

In crosssection 4.1.4 on p. 34 I have treated ostensible Purpose satellites which are to be described as Disjuncts. They indicate the intention of the speaker with his statement. A different kind of Purpose satellite is found in:

(66) Decii corpus ne eo die inveniretur, nox quaerentes oppressit (`The body of Decius could not be found that day, for night overtook the searchers', Liv. 8.10.10) 28a

(67) inde L. Genucio et Q. Servilio consulibus et ab seditione et a bello quietis rebus ne quando a metu ac periculis vacarent, pestilentia ingens orta (`Then during the consulate of L.G. and Q.S., when the situation was quiet and free from revolts and war, an enormous pestilence broke out, lest they ever be deprived of fear or dangers', Liv. 7.1.7)

For - rare - examples see K.–St. (II. 251, A.4). Utterances of this kind do not express the intention of an argument of the main predication, as in the case of the real Purpose satellites cited above, but with hindsight the speaker postulates a relationship of intentionality between main predication and embedded predication, suggesting an intervention by higher powers. [29]

(x) Result

(68) si quando non pluet ut terra sitiat (`If at some point it will not rain, so that, as a result, the ground becomes thirsty', Cato Agr. 151.4)

-- 121 --

Subordinator: ut (non) (`so that (not)').

Correlating expressions: ita, sic, adeo, usque eo (`so much'), etc.

For examples see K.–St. II. 247 ff. [30]

(xi) Condition

(69) si Fabius oriente Canicula natus est, Fabius in mari non morietur (`If F. was born while the dog-star was rising, F. will not die at sea', Cic. Fat. 12)

(70) si venisses ad exercitum, a tribunis militaribus visus esses (`If you had come to the army, you would have been seen by the army tribunes', Cic. Inv. 1.87)

Subordinator: si (`if').

Correlating expressions: sic, ita (`so'), ea lege (`on that condition'), idcirco (`therefore').

For examples see K.–St. II. 387 ff. [31]

7.3.2 Embedded predications as satellites in the function Disjunct

Instances of such satellites we have seen in chapter 4 (pseudo-Purpose satellites ( crosssection 4.1.4., on p. 34), pseudo-Conditions ( crosssection 4.1.5., on p. 35), pseudo-Cause satellites ( crosssection 4.1.6., on p. 36), limitation of validity ( crosssection 4.1.7., on p. 36)).

7.3.3 Clauses with or without a correlating adverb

In crosssection 6.5.1. (on p. 90) we briefly touched upon the difficulty of the description of the relation between a relative clause and a correlating pronoun such as is. The same difficulty arises in describing the relation between the ubi-clause in (45) and ibi in the main sentence.

(45) exsilium ibi esse putat, ubi virtuti non sit locus (`He thinks that exile is there, where there is no room for virtue', Cic. Mil. 101)

Just as in the case of relative clauses, there are clear instances in which the subordinate clause is Attribute ((71)) or independently functions as a satellite within the sentence ((72)):

(71) homines tenues … adeunt ad ea loca, quae numquam antea viderunt, ubi neque noti esse eis quo venerunt … possunt (`Poor people go to places which they have never seen before, where they cannot be known to those to whom they have come', Cic. Ver. 5.167)

(72) ut mulier … nihil putaret agi callide posse ubi non adesset Aebutius (`So that the woman thought that no clever deals were possible while A. was absent', Cic. Caec. 13)

-- 122 --

In (71) the ubi-clause can be seen as Attribute with ea loca in the main sentence, as also the preceding relative clause quae … viderunt is an Attribute with ea loca. In the subordinate clause ubi is a relative adverb [32] in the function Place Adjunct. Similarly, quae is Object in the relative clause in this example (see p. 82). In the subordinate clause in (72) ubi is also a relative adverb in the function Place Adjunct. The subordinate clause as a whole functions as a Place Adjunct within the main sentence. In a similar way, relative clauses may have an independent function within the sentence (see crosssection 6.5.1. on p. 90). The grammars tend to reconstruct independent subordinate clauses with a relative pronoun or adverb into clauses with an explicit antecedent. (72) would, for instance, be equated with (72a):

(72a) ut mulier nihil putaret agi callide posse ibi, ubi non adesset Aebutius

In a case like (72) the grammars speak of ubi `with implied antecedent'. Also, they usually equate the relation between the ubi-clause and the preceding ibi in (45) with the relation between the ubi-clause and ea loca in (71), which is modified by the subordinate clause. However, just as Attributes with is do not mean very much to us (is bonus (`the good he') does not exist), Attributes with ibi do not seem very plausible either.

The function of a number of subordinators within the subordinate clause is less easy to determine, e.g. si. Also, for si there are no clear correlating adverbs, as there are for example, for ubi.

7.4. Discussion of some individual constructions

This section deals with a number of the constructions mentioned in crosssection 7.2. and crosssection 7.3: first of all, in crosssection 7.4.1 a number of criteria which serve to distinguish embedded predications functioning as arguments and embedded predications functioning as satellites. In crosssection 7.4.2 I discuss the differences between the AcI and the construction consisting of accusative + prolative infinitive. In crosssection 7.4.3 I deal with predicates which allow both constructions. crosssection 7.4.4 concerns the interchangeability of the infinitive construction and ut-clauses. In crosssection 7.4.5 I discuss the Nominativus cum Infinitivo (NcI), and in crosssection 7.4.6 the so-called Accusativus cum Participio (AcP). In crosssection 7.4.7 the Dominant participle construction is treated. In crosssection 7.4.8 I recapitulate a number of restrictions on embedded predications.

7.4.1 Criteria to distinguish between arguments and satellites

A number of grammars, e.g. K.–St. (II. 272; 208; 232), make a distinction between final (i.e. Purpose) `noun clauses', i.e. arguments, and final `adverb clauses', i.e. satellites. Examples are (73) and (74), respectively:

(73) quod me ut scribam aliquid hortaris (`As to the fact that you incite me to write something', Cic. Att. 2.12.3)

-- 123 --

(74) maiores nostri ab aratro adduxerunt Cincinnatum illum ut dictator esset (`Our forefathers got the famous C. from behind the plough in order that he become dictator', Cic. Fin. 2.12)

Other, largely historically oriented grammars such as Sz. do not make the distinction noun clauses: adverb clauses, but simply speak of `Purpose clauses'. The equation or combined treatment of constructions of this kind has several causes.

(a) Formal similarities

There is a formal similarity between the two types of embedded predication, in spite of their different functions in the sentence and the fact that they are not semantically equivalent. The Complement clause introduced by ut in (73) does not represent the purpose of the exhortation, but its content. In (74), on the other hand, the ut-clause does indicate the Purpose of the state of affairs of the main predication.

(b) Nominal equivalents

Some constructions occurring with the predicate hortari are adduced to prove that the Complement of hortari in fact fulfils the semantic function Purpose, e.g. (75):

(75) ipsum tamen Pompeium separatim ad concordiam hortabor (`Yet, I will encourage P. himself separately to adopt a concordant attitude', Cic. Att. 7.3.5)

In many grammars we do not find a clear distinction between the use of prepositions to mark obligatory constituents and their use to mark satellites. As a result, instances like (75) are interpreted as Purpose Adjunct, on the basis of the presence of the so-called `final' ad. (For `final' ad in satellites see examples (64) and (66) above). In reality, ad concordiam in (75) indicates the content of the exhortation. There are also unequivocal instances, which impose the conclusion that the ut-clause with hortari is an argument and fulfils the syntactic function Complement. Examples are:

(76) sin tu, quod te iam dudum hortor, exieris (`If, as I have been encouraging you to do for a long time, you have gone away', Cic. Catil. 1.12)

(77) equidem pacem hortari non desino (`I do not stop urging for peace', Cic. Att. 7.14.3) [33]

(78) hortantibus dehinc quibusdam inediam et lenem exitum (`While some were offering the advice of a gradual death as a result of starvation', Tac. Ann. 11.3.2)

From these examples it becomes clear that hortari can also govern nominal constituents in the function Complement, and that an anaphoric pronoun in the accusative may be used to refer to the embedded predication.

-- 124 --

(c) Restrictions on tense

Satellite and argument ut-clauses with verbs such as hortari share at least one restriction: the tense of the subordinate clause cannot be anterior to the tense of the main sentence. The following instances would be impossible:

(73') *quod me ut scriberem aliquid hortaris (`* As to the fact that you incite me to have written something')

(74') *maiores nostri ab aratro adduxerunt Cincinnatum illum, ut dictator fuisset (`*Our forefathers got the famous C. from behind the plough in order that he had become dictator')

Besides similarities between argument and satellite ut-clauses such as those discussed here, there are also differences.

(a) Omissibility

A fundamental difference between the two constructions is, of course, the fact that argument clauses with hortari cannot be omitted, whereas satellite clauses can.

(b) Question words

A second difference, which was already mentioned above, lies in the fact that satellite ut-clauses provide the answer to a question cur?/quo consilio?, whereas argument ut-clauses do not.

(79) tene me, obsecro :: Cur? :: ne cadam (`Hold me, please :: Why? :: In order that I do not fall', Pl. Mil. 1260)

(c) Correlating expressions

In crosssection 7.3.1. (ix) (p. 119) we have seen that satellite ut-clauses may correlate with an adverb (idcirco) or another expression with a similar function (e.g. eo animo) in the main sentence. Argument clauses cannot.

(d) Juxtaposition/coordination

In crosssection 7.3.1. (viii) (p. 119) we have seen that satellite ut-clauses can be coordinated with other satellite clauses. This possibility of coordination usually entails identity of semantic and syntactic function (see above p. 30). Juxtaposition, on the other hand, proves that there is a difference in semantic and syntactic function. Example (80) shows that the argument clause (ut … consequare) is juxtaposed to the ut-satellite (ut … uteremur).

(80) intellexi … nihil mihi optatius cadere posse quam ut tu me quam primum consequare ut … tuo tuorumque praesidio uteremur (`I understand that nothing better could happen to me than for you to join me as soon as possible, so that I can make use of the protection offered by you and your companions', Cic. Att. 3.1) [34]

-- 125 --

(e) Referential identity

Argument ut-clauses with three-place predicates as a rule involve referential identity between one of the three arguments of the main predication and one of the arguments of the embedded predication. For an example of a three-place predicate see (43a):

(43a) patri persuasi, ut aes alienum filii dissolveret

The Addressee (patri) is referentially identical to the Agent of dissolveret. [35] In the case of two-place predicates, there is either identity of the first arguments of the two predications, or there is no identity at all, see (29c) and (29a), respectively:

(29c) virtutum in alia alius mavult excellere

(29a) tu malim … actum ne agas

Satellite ut-clauses do not necessarily require such referential identity, see (81):

(81) Caesar singulis legionibus singulos legatos et quaestorem praefecit uti eos testes suae quisque virtutis haberet (`C. gave each legion a legate and a quaestor as leaders, in order for everyone to have in them witnesses of their own courage', Caes. Gal. 1.52.1)

Satellite ut-clauses are naturally also found with one-place predicates, see (82):

(82) eo … ad te animo venimus, ut de re publica esset silentium (`We have come to you with the following intention, viz. that there be no talk of politics', Cic. Brut. 11)

(f) Controllability

In the case of argument ut-clauses with certain classes of verbs (e.g. hortari) the embedded predication has to be controllable. This is relatively obvious: it is no use exhorting someone to do something which is not in his power. This means, for instance, that the Addressee of the main sentence may not fulfil the semantic function Patient in the embedded predication and that this embedded predication may not be a state. Cf. in contrast (74).

(g) Omissibility of the conjunction

The two types of construction furthermore differ in that in argument ut-clauses the subordinator (ut (ne)) is omissible (type fac venias (`make sure that you come')), something which is impossible in `real' Purpose clauses. [36]

We see, thus, that there are four general criteria (viz. (a)-(d)) [37] to prove the difference between the constructions, while there are also specific criteria to bring to light the differences in the case of verbs of communication and emotion. Particularly (e) and (f) show that the constructions merely seem to be the same. In reality there appear to be semantic restrictions on the content of

-- 126 --

this construction if it is used as an embedded predication with e.g. hortari. We have already seen such restrictions in crosssection 2.4., and will again in what follows.

7.4.2 AcI and accusative + prolative infinitive

Strictly speaking, the term `Accusativus cum Infinitivo' (AcI) is used for constructions such as:

(83) dico te venire (`I say that you come')

In this example te venire as a whole is an argument containing the content of the statement (in the syntactic function Object with dico). [38] Te in (83) is not the person to whom the statement is addressed, since such an Addressee with dicere is marked by the dative:

(84) dico ei te venire (`I say to him that you come')

Often, however, the term AcI is used more broadly, so as to include cases like (85): [39]

(85) admoneo te venire (`I warn you to come')

Formally, this is, of course, an instance of an accusative (te) and an infinitive (venire). In contrast to (84), however, here it is not correct to regard te venire as a whole as an argument (in the function Object), since admonere is a three-place verb with one argument referring to the person who is warned (semantic function Addressee, in the active voice syntactic function Object) and one argument referring to the content of the warning (with the function Complement). [modified 12-08: Cf. (86a,b) and (87) for nominal and clausal expression of the Complement, respectively:

(86a) eam rem nos locus admonuit (`The place reminded us of the event', Sal. Jug. 79.1) [40]

(86b) de quo (proelio) vos … admonui (`Of which I have informed you', Cic. Man. 45)

(87) Antipater … admonere reliquos potuit ut adcuratius scriberent (`Antipater might have served as a warning to his successors that they should take greater pains at writing', Cic. Leg. 1.6)]

Infinitives such as venire in (85) will from now on be called `prolative infinitives' (for admonere see also crosssection 7.4.3 on p. 128). A comparison of the embedded predications and noun phrases shows that differences in semantic and syntactic structure are hidden beneath the ostensibly identical forms. In the remainder of this section I discuss a number of differences between the `real' AcI (example (83)) and the `accusative + prolative infinitive' (example (85)). For verbs which allow two constructions (accusative + prolative infinitive and accusative + AcI) see below, crosssection 7.4.3.

(a) Addibility of Addressee

Above it has already been pointed out that a constituent in the semantic function Addressee can be added to dico te venire. In the case of admoneo

-- 127 --

te venire this is impossible, since here te is the Addressee (but see crosssection 7.4.3 below).

(b) Restrictions on the predicate and on the arguments of the embedded predication

With admonere the warning cannot refer to something which has already happened prior to the warning. Consequently, we do not find:

(88) *admoneo te venisse (`*I warn you to have come')

We do, on the other hand, find instances such as:

(89) dico te venisse (`I say that you have come')

The AcI with dicere is not subject to restrictions as to the controllability of the embedded predication, or as to the animacy of the accusative constituent. With dicere we find e.g.:

(90) quid … spectans deus ipse diceret Marcellum … in mari esse periturum (`What could the god himself be aiming at in saying that M. would die at sea', Cic. Fat. 33)

(91) in quo iudicio … de verbis quaesitum esse dicatur (`A trial which, as will be argued, was about words', Cic. Caec. 38)

(92) quod (Xenophon) diceret eosdem labores non aeque gravis esse imperatori et militi (`Because X. said that the same efforts are not equally strenuous for a general and for a soldier', Cic. Tusc. 2.62)

In (90) we find a non-controllable predicate (perire); in (91) the passive of a controllable predicate; in (92) the Subject of the embedded predication is an inanimate entity. In embedded predications with admonere, on the other hand, there are, in fact, restrictions on these two points (controllable, animate). [41]

(c) Substitution

In the case of admonere the infinitive [modified 12-08: corresponds to, and can be replaced by,] a finite subordinate clause with or without the subordinator ut (ne). The constituent in the accusative – in the active voice - is still the entity that is being warned. The two constructions are synonymous: [42]

(93) illud me praeclare admones …, ne nimis indulgenter et ut cum gravitate potius loquar (`You rightly admonish me to do this, viz. not to speak too leniently, but rather with some dignity', Cic. Att. 9.9.2) [43]

The infinitive of dicere + AcI cannot be replaced (see p. 129). The criteria mentioned here clearly show the difference between the AcI proper and the construction of accusative + prolative infinitive. [44] We will see a similarity in crosssection 7.4.5. on p. 130 about the NcI.

-- 128 --

7.4.3 Predicates allowing both AcI and accusative + prolative infinitive

The description of admonere in reality turns out to be slightly more complicated. With admonere we not only find a three-place construction with a prolative infinitive as third argument (like admoneo te venire) but also a three-place construction with an AcI-construction as third argument (like admoneo te hostes venire, `I warn you that the enemies are coming'). In English, too, these frames may be discerned `to warn someone to … ' and `to warn someone that something is the case', respectively. Examples of admonere aliquem + AcI are:

(44d) tantum te admonebo … te his daturum (salutem) (see p. 115)

(44e) admonitus sum ab illo … dici posse (see p. 115)

This `true' AcI is naturally not subject to the restrictions mentioned above ( crosssection 7.4.2.) for admonere + accusative + prolative infinitive.

The same possibility (three-place construction with as third argument either a prolative infinitive or an AcI) is found with docere (`to teach'), since this verb, too, has an Addressee marked by the accusative (in the active voice); see examples (39e) and (39d) on p. 113. Other verbs of communication, such as persuadere, do not show this at first sight striking co-occurrence of two accusative constituents with one predicate, since there the Addressee is as a rule marked by the dative (see example (43d) on p. 115: hos homines cannot be interpreted as Addressee). The difference between prolative infinitive and AcI with admonere and docere correlates with the difference between imperative and declarative modality of the embedded predication. We return to this difference in crosssection 7.4.4.

The verbs meaning `to order', in particular iubere, constitute a separate problem, which did not appear clearly from tables 7.4 and 7.5 on pp. 108 and 112, respectively. [45] With this group of verbs we find both a three-place construction with a prolative infinitive and a two-place construction with a `true' AcI. Examples of the three-place construction are (41d) and (42c) on p. 114. Examples of the two-place construction referred to are:

iussi...ianuam claudi (see (42d))
/...ianuam patere
imperavi...te laudari

Also possible is (two-place!!):

iussi...te venire
imperavi...te venire
(`I have given the order that you must come')

In the interpretation aimed at here te is not the Addressee of the order. Naturally, iussi te venire can also mean `I have ordered you to come'; this is,

-- 129 --

therefore, a case of structural ambiguity. Note that this ambiguity does not exist in the case of imperare, since with imperare the Addressee is marked by the dative (imperavi tibi venire). [46]

7.4.4 Interchangeability of infinitive construction and ut-clause

With the predicate dicere we find besides the AcI construction also embedded predications in the form of an ut-clause, where the ut-clause in its turn may be replaced by a prolative infinitive (see K.–St. I. 683).

(38b) dices (eis) … paulum proferant auctionem

(38b') dices (eis) auctionem proferre [47]

The ut-clause and the prolative infinitive construction with dicere are subject to the same restrictions that we saw above for the prolative infinitive with admonere and the ut-clause with hortari: the embedded predication must be `controlled', and cannot, therefore, be passive; anteriority of the infinitive is impossible; in the embedded predications expressions like nimirum, fortasse, haud dubie are excluded. [48] The modality of the embedded predication is imperative (see crosssection 7.2.3.), whereas this modality is declarative in the case of the AcI. Other predicates which allow embedded predications with declarative or imperative modality are the verbs of communication, respondere (`to answer'), scribere (`to write'), clamare (`to shout'), nuntiare (`to report'), etc. We also find this alternation with a number of two-place verba sentiendi: decernere (`to determine'), statuere (`to decide'), cogitare (`to consider'), etc. [49] As was stated above, admonere also belongs to the class of verbs that allow both an embedded predication with declarative modality (= admonere aliquem + AcI) and an embedded predication with imperative modality (= admonere aliquem + ut-clause or prolative infinitive). Schematically, this may be represented as table 7.6. Table 7.6
Predicate (incl. Subject)Addresseecontent
dicotibime abireabire/ut abeas
admoneoteme abireabire/ut abeas
statuoØme abire (= to determine that)abire/ut abeam (= to decide to)

From this table it appears that the ut-clause and the prolative infinitive construction are synonymous and that they are in opposition to the AcI-construction (but see above crosssection 7.4.3. about iubere). Incidentally, the verbs with

-- 130 --

which ut-clause and prolative infinitive constructions are or ought to be possible differ in the extent to which we actually find the two constructions. There is, especially in poetry (see K.–St. I.680), a general tendency in favour of the infinitive construction. [50]

7.4.5 Nominativus cum Infinitivo (NcI)

In crosssection 7.4.2. we have seen that in dico te venire the constituent te has no semantic relation with dico, but that te venire forms a whole that with an active main predicate fulfils the syntactic function Object. This also becomes apparent in the passive: dico te venire can be transformed into:

(96) dicitur (a me) te venire (`It is said (by me) that you come')


(97) ei … dictum est clipeum esse salvum (`It has been said to him that his shield was safe', Cic. Fam. 5.12.5)

(98) dicitur eo tempore matrem Pausaniae vixisse (`It is said that the mother of P. lived in that period', Nep. Paus. 5.3) [51]

With admonet te venire, on the other hand, there is a semantic relation, te functioning as Addressee and fulfilling in the active voice the syntactic function Object. Consequently, passivization is possible:

(99) admoneor venire (`I am warned to come')


(100) vos … admonendos puto ne … putetis (`I think that you should be warned not to think', Cic. de Orat. 3.201)

Now, a problem is posed by instances such as:

(101) dicor venire (`I am said to come')

(102) eruptionem facturi fuisse dicebantur (`They were said to be going to break out', Cic. Att. 7.14.2)

This so-called personal construction is more common than the `impersonal' construction exemplified above. In spite of the fact that in (101) I is semantically speaking not an argument of dicere but of venire, I nevertheless becomes (syntactically speaking) Subject of the passive form of dicere. An argument of the embedded predication is, as it were, promoted to the main predication (see further crosssection 7.5. on p. 135 ff.). [52] On the other hand, (101) cannot be equated with (99). It turns out that the restrictions mentioned on p. 124 ff. are also applicable in the passive, e.g. the restriction on the tense of the embedded predication:

-- 131 --

(99a) ?admoneor venisse [53] (`I am warned to have come')

While the following expression is possible:

(101a) dicor venisse (`I am said to have come')

In explaining the occurrence of the personal passive constructions with dicere most grammars tacitly assume that the personal construction and the impersonal construction are synonymous. K.–St. I. 707–9, however, already point out that under certain circumstances the impersonal construction is preferred, e.g. when the point at issue is to whom, why or how a statement is made. The personal construction, on the other hand, turns out to be preferred if a constituent in the embedded predication is Focus and the embedded predication as a whole does not constitute a pragmatic unit (e.g. in the case of a question as to `who has been told to do something'). There are, therefore, pragmatic differences between the two constructions. See also crosssection 12.3.3. [54]

A completely different type is represented by the following, mainly poetic, constructions (see K.–St. I. 702).

(38f) quas hodie adulescens Diabolus ipsi daturus dixit

(103) sensit medios delapsus in hostes (`He noticed that he had ended up in the midst of the enemies', Verg. A. 2.377) [55]

7.4.6 Accusativus cum Participio (praesentis)

The AcP-construction, as exemplified in (26f) and (104), occurs in Classical Latin [56] with verbs of `perception'.

(26f) idque Socratem … audio dicentem

(104) adulescentium greges Lacedaemone vidimus ipsi incredibili contentione certantes pugnis (`We have seen with our own eyes troops of youngsters in Lacedaemon fighting with inconceivable obstinacy, using fists', Cic. Tusc. 5.77) [added 12-08] Besides the AcP this group of verbs also allows the AcI (the reverse is not always the case!). The difference between the two constructions is that in the case of the AcP the aspect of `perception' is central, and with the AcI that of `cognition' and `reflection' (see Sz. 387; Bolkestein 1976b: 283–8). In practically all instances of the AcP the accusative constituent can also be interpreted as an argument with the verb of perception. In example (26f), for instance, it can be argued that Socrates is heard while he is speaking. The only instance known to me that might be considered an exception is (105): [57]

(105) at ille ut Carthaginem venit multo aliter ac sperarat rem publicam se habentem cognovit (`But when he came to Carthage, he noticed that the political situation was quite different from what he had expected', Nep. Ham. 2.1)

K.–St. (I. 704) interpret this instance as an example of an AcP. In this case, the meaning must be that he could perceive the changes with his own eyes. [58] Against this analysis it can be objected that with two-place cognoscere in this sense we do not find nouns such as res publica as Object. It might be better to

-- 132 --

regard rem publicam and se habentem as Object and Object Complement, respectively (see p. 22), since there are parallels for three-place cognoscere (`to find (to be)':

(106) firmitatem … et constantiam … eandem cognosces quam reliquisti (`You will find my firmness and constancy to be the same as when you left', Cic. Fam. 9.11.1)

Sz. treats the AcP as a subcategory of the use of the participle as Praedicativum. It is true that in many cases omission of the participle and the constituents governed by it does not result in ungrammaticality of the remainder of the sentence. [59] In example (104) above this is, of course, impossible.

K.–St. I.704 classify in this group also instances such as (107), where no perception is involved. I consider these instances as clear examples of participles in the function Praedicativum. Unlike (26f), here replacement with an infinitive construction is impossible. 59a

(107) offendi eum sedentem in exedra et cum C. Velleio senatore disputantem (`I came upon him seated in the sitting room, in a discussion with the senator C.V.', Cic. N.D. 1.15)

7.4.7 Dominant participle construction

Examples of the so-called dominant participle [60] have been given in (19b) and (30e):

(19b) occisus dictator … pulcherrimum facinus videbatur

(30e) cum … valde absoluto Scaevola gauderet

In (19b) the constituent occisus dictator is an argument in the syntactic function Subject; in (30e) the constituent absoluto Scaevola is also an argument, but here in the syntactic function Complement. In crosssection 7.3.1 (vii) `Accompanying circumstances' (on p. 117) we have given examples of dominant participle constructions as satellites (the so-called ablative absolute): 60a

(52) (Cethegus) recitatis litteris … repente conticuit

-- 133 --

We have also seen the dominant participle construction on the noun phrase level (p. 79). Examples are:

(108) suspicio acceptae pecuniae (`The suspicion of having received money', Cic. Ver. 38)

(109) ante conditam … urbem (`Before the foundation of the city', Liv. 1, pr. 6) [61]

The dominant participle construction differs both from constructions in which the participle occurs as Praedicativum (see crosssection 8.5.5. on p. 160) and from constructions in which the participle occurs as Attribute. In the dominant participle construction the participle cannot be omitted. The result of omission of the participle from (19b), for instance, is ungrammatical:

(19b') * dictator … pulcherrimum facinus videbatur

The non-omissibility can also be illustrated with the aid of an instance such as (50). The sentence which remains after imperatore has been omitted is ungrammatical: [62]

(50) * Bellum Gallicum C. Caesare gestum est

As becomes clear from the series of examples of gaudere in (30), the dominant participle construction is merely one of the possible forms of an embedded predication. Alternatives for example (19b) cited above might be:

(19b") quod dictator occisus erat, pulcherrimum facinus videbatur

(19b"') dictatorem occisum esse pulcherrimum facinus videbatur

In many cases the dominant participle construction may also be replaced by a noun phrase consisting of an action noun and a noun phrase in the genitive:

(19b"") caedes dictatoris pulcherrimum facinus videbatur

Both the non-omissibility of the participle and the fact that the construction as a whole can be replaced by alternative constructions show that the construction must be regarded as one whole.

Above I have shown how the dominant participle construction may be replaced by alternative constructions. It is, however, not the case that these alternatives are equivalent and mutually interchangeable. A dominant participle construction functioning as argument occurs only if it is implied that the state of affairs referred to by the dominant participle construction has in fact been realized, is being realized or will be realized, in other words, if the event referred to is `factive'. [63] Consequently, an embedded predication in the form of a dominant participle construction can only occur as argument with certain, `factive', main predications; we do, for instance, find such an embedded predication with pulcherrimum facinus videbatur, but not with difficilis esse:

(19b""') * Caesar occisus difficilis fuit

-- 134 --

There are no restrictions on the noun phrases and the predicates which may occur in the dominant participle construction itself. Also, there are no restrictions on the tense of the participle. An example of a dominant participle construction with a present participle is (110): [64]

(110) fugiens … Pompeius … homines movet (`The fact that P. is fleeing shocks the people', Cic. Att. 7.11.4)

The Predicate of a dominant participle construction need not necessarily be a verb, but may also be an adjective or a noun. Examples are (51), (111) and (112), respectively:

(51) qui tranquillo mari gubernare se negent posse (`Who say they cannot sail on a calm sea', Cic. Rep. 1.11)

(111) augebat metum gnarus Romanae seditionis et, si omitteretur ripa, invasurus hostis (`The fear was increased by the fact that the enemy was aware of the rebellion among the Romans and would make an invasion if the river bank were no longer guarded', Tac. Ann. 1.36.2)

(112) filius legati orator publicae causae satis ostenderet necessitate expressa quae per modestiam non obtinuissent (`The fact that the son of a legate acted as champion of the public cause made clear that they had exacted with force that which they had not been able to obtain with modest behaviour', Tac. Ann. 1.19.5)

7.4.8 Survey of the restrictions and criteria relevant to complex sentences

In this section I review a number of important restrictions treated in this chapter:

(i) Addibility of Disjuncts that limit the `truth value' of a sentence to certain sentence types (fortasse etc.: excluded with imperative modality (p. 101; p. 129)).

(ii) Factivity: the use of the dominant participle construction in embedded predications is restricted to those main predicates which presuppose that the event has taken place, is taking place or will take place. On the noun phrase level the dominant participle competes with the (non-factive) gerundive (p. 79; p. 133).

(iii) Referential identity: often there is referential identity between arguments of the main predication and the embedded predictaion (e.g. with persuadere alicui). In cases of referential identity the argument concerned may be omitted in the embedded predication. (In transformational-generative grammar this is called `Equi-NP-deletion') (p. 125 f.).

(iv) Tense: in embedded predications with verbs meaning `to order', `to wish', etc. the state of affairs referred to by the embedded predication must be posterior to the moment referred to by the main predication.

-- 135 --

(v) Controllability: embedded predications with verbs meaning `to order' must be controllable; this becomes clear e.g. in the restriction on the tense of the embedded predication mentioned in (iv) and in the impossibility of referential identity between the Addressee of the main predication and the Patient of the embedded predication.

7.5. Personal and impersonal constructions

While discussing embedded predications with constat and difficile est (p. 104 and p. 106, respectively), I referred to the personal constructions (10b) and (13e). crosssection 7.4.5. dealt with the NcI, which may be considered a personal counterpart of constructions such as dicitur + AcI. In this section I deal with the relation between personal and impersonal constructions. Besides the phenomena mentioned above, I also discuss the verbs debere and posse.

7.5.1 Constat

TLL s.v. consto 535.49 calls (10b) an example of an NcI:

(10b) quae si omnia e Ti. Coruncanii scientia … acta esse constarent

I reserve the term NcI for the passive counterparts of the AcI, as treated in crosssection 7.4.5. There is one other Classical instance, Cic. Clu. 104 (see also note 4). There are personal constructions of constare in the sense `to be established, well-known', e.g. (113) and (114):

(113) eorum … quae constant exempla ponemus (`We will give examples of those things that are established', Cic. Inv. 1.68)

(114) cum et factum constat et nomen (`When both the fact and its name are clear', Cic. Part. 42)

Example (10b) cannot very well be viewed in relation with these instances. Also, that which is established in (10b) is not omnia but the state of affairs omnia … acta esse. There is, therefore, as in the case of the NcI, a discrepancy

Figure 7.1

-- 136 --

between the semantic structure and the syntactic structure; this may be represented as in figure 7.1. The Subject of the embedded predication is, as it were, promoted to Subject of the main predicate. In transformational terms one might speak of `Subject-to-Subject raising'. [65]

7.5.2 Copula + adjective + supine

Above (p. 106) I have treated (13e) as an instance as an example of an embedded predication, together with the impersonal constructions.

(13e) difficile est hoc genus exornationis inventu

This is because here, too, difficilis is not predicated of the entity genus exornationis but of the state of affairs invenire genus exornationis. This may be represented as in figure 7.2.

Figure 7.2

This figure indicates that the Object constituent [66] of invenire is promoted to Subject of the main predicate. A parallel is the following instance, much discussed in transformational grammar:

(115a) John is easy to please

(115b) It is easy to please John

The construction of adjective and supine (in -u) [67] is found with the following classes of adjectives (see K.–St. I. 724):

(a) (i) physical characteristics, e.g. asper (`rough'), foedus (`ugly'), pulcher (`beautiful');

(ii) value judgments, e.g. crudelis (`cruel'), honestus (`honourable'), turpis (`shameful');

(b) adjectives meaning `difficult', `easy', `possible', e.g. difficilis (`difficult'), facilis (`easy'), incredibilis (`unbelievable'), mirabilis (`surprising').

-- 137 --

Among the instances mentioned in the literature there are relatively few cases in which the copula esse is present. Most of the instances occur on the noun phrase level. Especially in poetry the use of adjective + supine on the noun phrase level is productive (in particular with dictu). [68]

For a number of adjectives, especially those of class (a), the treatment given above for difficilis is not satisfactory. Thus, (116a) is not synonymous with (116b):

(116a) o rem … visu foedam (`A matter, terrible to see', Cic. Phil. 2.63) [69]

(116b) foedum est istam rem videre

What is terrible is not seeing the matter, but the matter itself. To my mind, of (117a–b) it is slightly less evident that the two expressions are not interchangeable:

(117a) omnia praeteribo quae mihi turpia dictu videbuntur (`I will leave aside all those things that seem to me shameful to say', Cic. Ver. 1.32)

(117b) turpe mihi videtur ista omnia dicere

The facts are shameful, but speaking about them need not necessarily be shameful. In view of the fact that it is difficult generally to apply the description given for difficilis one might consider not to treat the construction copula + adjective + supine in terms of embedded predications, but to regard adjective + supine as a complex unit. [70]

7.5.3 NcI

We have also encountered a discrepancy between syntactic and semantic structure in crosssection 7.4.5. Example (101) could be represented as in figure 7.3.

Figure 7.3

In transformational terms, this is an instance of `Subject-to-Subject raising'.

7.5.4 Debere and posse

At first sight, debere (`must, have to') [71] seems to have a structure similar to that of velle (`want'):

-- 138 --

(118) debeo ire (`I must go')

(119) volo ire (`I want to go')

In other words, a two-place predicate with a Subject and a Complement. From a semantic point of view, it is less attractive to assume a similar structure for (118) and (119). Velle involves a relation of will between someone who wants something (Subject) and something that is wanted (Complement), the element of will being located in the former. Consequently, the Subject constituent is subject to semantic restrictions (must be human). Debere, on the other hand, is more adequately described as containing a relation of obligation between Subject and Complement where the element of obligation is not located in the former. The obligation may be due to external persons or entities. Debeo ire may be described as `it is obligatory, necessary, that I go', comparable to oportet me ire (`I ought to go'). If we assume the same semantic structure for debere as for oportet, the Subject I, semantically an argument of the embedded predication, fulfils the syntactic function Subject with the main predicate (so: Subject-to-Subject raising). [72] This leads to the representation as in figure 7.4.

Figure 7.4

Also in the case of posse in the sense `be possible' [73] a `Raising' analysis might be attractive. An example is (120):

(120) cum haec scribebam V Kalend., Pompeius iam Brundisium venisse poterat (`While I am writing this (25 February), it is possible that P. has already arrived in B.', Cic. Att. 8.9A.2)

What is at stake here, is not so much Pompeius' capacities as the possibility that something has happened. 73a

7.6. Historical approach to the phenomenon of complex sentences and of certain types of complex sentence

In this section I first discuss current opinions in Latin linguistics concerning

-- 139 --

the origin of complex sentences ( crosssection 7.6.1.), then the historical explanations offered for AcI and subordinate clauses introduced by a conjunction ( crosssection 7.6.2.).

7.6.1 Hypotaxis in general

All known languages, in spite of their considerable formal differences, have complex sentences, both those treated in this chapter and relative clauses. In view of this, it is not desirable to assume that earlier stages of Latin (e.g. Indo-European) did not allow the possibility of embedding predications in other predications. Yet, this is currently assumed, and in many cases complex sentences are explained as arising from two independent, simple sentences. [74] Such an assumption presumably arises from ideas on the development of the human brain, by analogy with the development of child language, regardless of the existence of empirical support (e.g. K.–St. II. 1). A second reason may be the fact that a number of subordinators can in all likelihood be shown to have developed from connectors [75] and adverbs, and it seems difficult to develop a hypothesis as to which subordinators existed in pre-historic periods. But the fact that we are unable to determine which specific subordinators existed does not mean that there were none. The Romance languages show no trace of highly common Latin subordinators such as ut and cum; if Latin had not been preserved, we would never be able to reconstruct these subordinators. General statements such as `In der ältesten indogermanischen Zeit gab es keine Nebensätze' (`In the oldest Indo-European period there were no subordinate clauses') are unfounded. [76]

7.6.2 Specific complex sentences

In this section I treat two types of complex sentence and their current historical explanations. These are (i) clauses introduced by a subordinator and (ii) the AcI.

(i) Clauses introduced by a subordinator

Scherer (1975) clearly illustrates the ways in which he thinks specific constructions have come into being. [77] Owing to this clarity, he offers a Table 7.7 Examples from Scherer (1975)
Latin sentenceTo be explained fromScherer
(a) ubi sim nescio`Where might I be? I do not know'p. 238
(b) num me vituperas? :: multum abest ut te vituperem`Are you criticizing me?' `Not at all; how could I criticize you?p. 253
(c) dux milites hortatur ut fortiter pugnent`… how they would fight courageously'p. 269
(d) nemo tam stultus est quin id intellegat`How could he not understand this? Nobody is that stupid'p. 271

-- 140 --

convenient point of departure for the discussion. I first present, in table 7.7, his examples and the constructions which he uses as his starting point. The types of explanation are evident. In (a) a paraphrase with two independent sentences is chosen, and it is suggested that this structure is also possible in Latin. However, an independent (deliberative) question ubi sim does not seem very plausible, and in the paraphrase either nescio is one-place or the second argument is to be understood from the context (and this second argument would, of course, be ubi sim). (b) is slightly more complicated: in the Latin sentence ut does not mean `how' and the speaker is not wondering `how he could criticize'. Here, too, multum abest would lack an argument. The analysis also presupposes a kind of conversation. These remarks with regard to (b) in part also hold for (c). In (d) the order must be manipulated; the order given in the English sentence is not (or rarely?) found in Latin.

In giving these examples I do not intend to prove that subordination could never have arisen from originally independent sentences; on the contrary, it is evident that examples may be found. But the postulated development often presupposes forms and constructions which deviate rather essentially from Latin as it is known to us, and a too ready use is made of seemingly plausible paraphrases in a modern language.

(ii) AcI

The `true' AcI is often regarded as identical with or at any rate as having originated from the accusative + prolative infinitive. Above (p. 126) I have shown that synchronically the two constructions are certainly not identical. Here we address the alleged diachronic relationship. Many linguists [78] assume a so-called `reanalysis':

(a) arguo praetorem : pecunias cepisse → arguo: praetorem pecunias cepisse

(b) nuntio eum: dictatorem esse factum → nuntio : eum dictatorem esse factum

An objection against this analysis is the fact that it suggests that in Classical Latin (the situation to the right of the arrows) (a) and (b) had the same structure; this was not the case. Moreover, in (b) it is suggested (to the left of the arrow) that besides the `reporter' at one stage another person could occur as argument with verbs such as nuntio, something which is highly unlikely from a semantic point of view. Furthermore, it is unclear what pecunias cepisse might have meant. [79] For arguere we must assume that the three-place construction current in Classical Latin (arguere aliquem stupri) did not exist. Here, too, the facts are strained unnecessarily. [80]

From this section it must be concluded that the existing hypotheses concerning the origin of complex sentences are too simple. This is not to say, of course, that I have as yet offered an alternative explanation for the situation as we find it.

-- 141 --

Bibliographical information

There is no recent study covering the whole area of complex sentences in Latin. Lambertz (1982) treats the greater part of the constructions, without, however, discussing the large number of specific treatments of more limited points. Lakoff (1968) offers a description of complex sentences in Latin in a transformational-generative framework which is relatively useless for Latin linguistics. Extensive bibliographical references for individual constructions are given in the notes. A very great number of publications concerns the AcI. 80a See for this Bolkestein (1976a; 1979). For various constructions with verba dicendi see also Bolkestein (1977b; 1977d; 1988b). For the NcI see Bolkestein (1980b; 1980c; 1981b; 1983b). The dominant participle (`ab urbe condita') construction and the ablative absolute are discussed in Bolkestein (1981b; 1983b) and Pinkster (1972b). For the supine see Kroon (1989b). For the function of subordinators see Pinkster (1972c: 174–8).

-- 142 --

Previous Section

Next Section

Pinkster, Harm (1942-) [1990], Latin Syntax and Semantics [info], xii, 320 p.: ill.; 24 cm. [word count] [Pinkster].
Powered by PhiloLogic