Greek Consecutive Clauses

‘When the result actually occurs, the result clause has an ordinary indicative verb.’ (GT2, p. 30)

‘result clauses … commonly have their verb in the indicative if the result is one that actually occurs.’ (GT2, p. 116)

(The italics are mine, as they are throughout unless otherwise indicated.)

The verb in a Greek consecutive clause may be in the indicative or infinitive. I don't believe that anybody has properly understood (as an ancient Greek would have understood, presumably) the difference of meaning signified by the indicative and infinitive, in particular the meaning(s) conveyed by the infinitive (though sometimes it seems as if there is no difference).Footnote ³ This has no doubt contributed to the misleading explanations that we find in the grammar and course books.

It seems to be supposed — it is actually asserted in some books (see the quotations from Taylor above) — that the indicative denotes a real or actual result as opposed to a ‘natural result’ (or some such expression) denoted by the infinitive. There is often an implication that the writer is showing by his use of the indicative that he knows that it was a real or actual result. This is a mistake. I am not sure what the use of the infinitive tells us — I don't think that anybody is sure. But I am sure (a) that the Greek infinitive cannot assert anything, just as the indicative when used in a statement cannot fail to assert something; (b) that we cannot tell from the use of the indicative alone that the result was a real or actual result, or that the writer knows that it was. For all we know it may not have happened. For all we know the writer may know or suspect that it did not happen. All we know for sure is that the writer says that it was an actual result.

The same is true of the writer's use of the infinitive — actually more common than the use of the indicative. Whatever kind of result it signifies, again all we can say is that the writer says that what is asserted in the main clause was such as to have a result of this kind.Footnote ⁴ Whether it did or not we have no way of knowing solely on the evidence of the way the writer expresses himself — nor, of course, whether the writer knew or didn't know.

Let us suppose that the infinitive is used to signify a natural, likely or to-be-expected result, and that the writer does not use the indicative because he does not wish to state that such a result actually occurred. The writer may be mistaken on both counts: such a result may have actually occurred, and it may have been a highly unnatural, unlikely and unexpected result. But we have no way of knowing whether it occurred and what type of result it was solely by the writer's choice of verb form. We do not know the facts ; we only know the writer's attitude to the facts , and this is all that his words convey.

Note too that the tense of the infinitive denotes aspect not time. The infinitive is a verbal noun, used to denote a type of action/occurrence, not an actual action/occurrence, and it cannot be used in Greek to assert anything. This can cause ambiguity about the time location (if we can talk of the time location of something that may not have occurred) of the consecutive clause. Since the tense of the infinitive is time-neutral, the time location of the expected/likely result cannot be shown by the form of the verb. How does Greek express unambiguously ‘He was so spendthrift that he is (likely to be) short of money’? As far as the grammar of the sentence is concerned, it could just as well mean ‘He was so spendthrift that he was (likely to be) short of money ‘. If context did not supply the (?correct) time location, time-specific adverbs or other aids to understanding could be used.

Greek Causal Clauses

Although the same constructions in Greek can be used for assigning causes to events and for ascribing reasons to agents, nobody (I think) seems to think it necessary to distinguish a cause from a reason, though ‘reason’ is not normally (it is sometimes, rightly or wrongly) used as a synonym of ‘cause’. (There is no agreement among philosophers as to what either of them is, except that they are not usually the same thing, or the same sort of thing; and there is disagreement about whether a reason may in fact be a cause of human agency.)

‘Causal clauses denoting a fact regularly take the indicative …’ (Smyth, A Greek Grammar, §2241). And in the next section he says ‘causal clauses denoting an alleged or reported reason’. There does seem to be the suggestion here of a distinction between an actual cause and an erroneously alleged cause: that we can tell what the actual cause was from the mood of the verb used.

When the main verb is in a historic tense (including the historic present) the verb in the causal clause may be in the optative rather than the indicative. This is because the clause is in effect an instance of ‘virtual oratio obliqua’: the introductory verb of saying or thinking is suppressed and must be ‘understood’ by the reader. This is the only reason for the optative. One should not read into the use of the optative an ‘alleged’ cause as opposed to an actual or real cause expressed by the indicative. In fact the opposite may be the case, and the writer may know or suspect that this is so. The use of the indicative does not indicate a real or actual cause and one known to be real or actual by the writer. The indicative should not be taken as a guarantee of actuality. It merely indicates a cause given by the writer rather than one attributed by the writer to someone else. As I said, the real cause may be the latter: the writer may be mistaken. But there is nothing mistaken about his choice of verb moods, and whether or not the cause was as stated is irrelevant to the writer's choice of mood. (Incidentally, the tense of the optative indicates time not aspect. And an original past tense (aorist) indicative assumed to have been used by the person to whom the cause is attributed may be represented by the aorist optative. This is an exception to the ‘rule’ that after a historic main verb a past tense indicative in a subordinate clause in direct speech must remain in the indicative in the corresponding subordinate clause in indirect speech (or, as here, virtual indirect speech).

If the main verb is in a primary tense, the optative may not be used, whatever the tense of the indicative in the assumed original causal clause. But the same distinction may be made by the writer, using other means. The writer may use ὡϛ (or ἅτε or oἱ̂α) with a participle to indicate an attributed cause, and a participle alone for a cause given by himself. Again, we should not regard the latter as a real or actual cause, and we should not suppose that the writer knows this to be so.

Causes may also be stated by means of a preposition such as διά or ἕνεκα and the articular infinitive: ‘because of/on account of X-ing’. The use of the infinitive, and the fact that the tense of the infinitive signifies aspect, not time, make it less likely that the verb form is taken as denoting an actual cause as opposed to one suggested by the writer.

A common way in an English translation to convey an attributed cause is to insert the words ‘as X said/thought’ as a parenthesis. This is infelicitous because ambiguous. It could be taken as an endorsement of what X said, i.e. ‘just as X said’, or as a simple reference to what X said, i.e. ‘according to what X said’, but perhaps with an implication that what he said was mistaken. One may end up with a translation that suggests the opposite of what the Greek says.

Greek Conditional Sentences

‘In the fifth sentence [a Past Unreal sentence] we are in the area of the unreal or impossible.’ (Morwood GG, p.183)

‘If X had happened, Y would have happened [Taylor's italics] (but X did not in fact happen, so Y did not happen either).’ (Taylor GT2, p.139). [my italics] See also in the same paragraph the statements about ‘known fact’ and ‘known not to be true’. (To be fair to Taylor, he does go on to say about the apodosis ‘would have (implying but in fact did not)’. The important word here is ‘implying’ (see below).

In the case of conditions the misunderstanding affects the whole sentence: the main clause (the apodosis) and the subordinate clause (the protasis). Certain types of conditional sentences are a minefield of misunderstanding.

It is difficult to talk about conditional sentences at all, because of the number of different names given to the different types of conditions, the same writer not infrequently giving different names to the same type of condition.Footnote ⁵ With apologies for adding to these, I use the names used by Woodcock in A New Latin Syntax , pp. 147-148, viz. ‘Open’, ‘Ideal’, ‘Unreal’ (the second two also used by Gildersleeve and Lodge, who curiously use ‘Logical’ for ‘Open’). It is important at the outset to understand the differences between the different types of condition. One must remember too that the protasis and apodosis may be in different tenses and be of different types. There is more scope for misunderstanding conditional sentences than consecutive or causal clauses. And three moods of the verb (indicative, subjunctive, optative) are involved rather than two (counting the infinitive here as a mood, which it is not, since, as I said earlier, it is a verbal noun).

Actually, the sort of misunderstanding of conditional sentences that I am concerned with here is more or less confined to the type of condition that I call ‘Unreal’. The ‘Open’ and ‘Ideal’ conditions do not cause the same problems. As far as Open conditions are concerned (those which use the indicative, or subjunctive for future open and present/future indefinite), nobody is tempted to mistake the supposition of a fact for a fact — the supposition of a fact is clearly different from the assertion of a fact. It is the assertion, or apparent assertion, of facts that get converted into facts. In the case of Ideal conditions (those which use the optative, also used for past indefinite) we are well removed from the realm of facts in any case, since these conditions are concerned with obviously hypothetical future contingencies. (Philosophers question whether one can talk meaningfully of future facts.)

It is the third type of condition, the ‘Unreal’ one, that causes the problems. These are of two kinds, Present Unreal of the form ‘If he were (now) doing X, he would (now) be doing Y’, and Past Unreal of the form ‘If he had done X, he would have done Y’. Both use the indicative, but are distinguished from Open Conditions by the presence of ‘n’ in the apodosis. They are called ‘Unreal’ because the possibility that the condition was fulfilled or is being fulfilled is treated by the writer as an unreal one — which does not rule out the possibility that the condition was in fact fulfilled or is being fulfilled. Again, the writer may be mistaken — or lying.Footnote ⁶

It is maintained erroneously that one can tell from the way in which the writer expresses himself that the opposite of what is supposed in the protasis and apododis was/is the case. First, it is said that the writer states that X and Y did not occur or is not now occurring. The writer does not state it, he implies it. The sentence can thus be unpacked as ‘If he had done X (it is implied that he did not in fact do X), he would have done Y (it is implied that he did not do Y)’. The error is to mistake an implication of a fact for a statement of a fact, and a statement of a fact for a fact. We cannot tell from the grammar of the sentence what the facts in the case are. The grammar of the sentence merely indicates the writer's attitude to the facts. The writer may be mistaken in his implication; he may know that he is mistaken in his implication. He may be lying. So, for all that we know, i.e. for all that we know from what the sentence actually tells us, our subject may have done X and he may have done Y. (Goodwin, followed by Dickey, calls Unreal conditions ‘contrary to fact’ or ‘contrafactual’ (Dickey), which is unfortunate since it gives the impression that the implication is a fact and that we can tell that the condition was not in fact fulfilled.)

‘If he had done that then he would have done wrong.’ Can we tell from the form of this conditional sentence that the condition was not fulfilled? Suppose this is a mother speaking about an errant offspring to an irate neighbour. She may know full well that he (bound to be a son, of course) did it, while implying that he did not do it, thus seeking to protect the offspring, and mollify the neighbour, by evasion (in the protasis) and agreement and acceptance of responsibility (in the apodosis). But again, all we have to go on is what the speaker actually says. The mother (and offspring) must hope that the neighbour is similarly in the dark. So here the opposite is the case from what the speaker implies. What the speaker implies is no guarantee of the actuality of the implication — and anyway an implication is not an assertion.

Latin Consecutive Clauses

Seemingly paradoxically, in certain instances the verb in the consecutive clause may be in a past tense while the verb in the main clause is in the present tense. The result clause may refer to present and future time as well as past time. So any tense of the subjunctive may be used, including the periphrastic future -urus sim and ‘future in the past’ subjunctive -urus essem — and the potential ‘would have …’ in - urus fuerim/fuissem.

The commonest tenses used, however, are the imperfect and perfect subjunctive. The distinction does not usually correspond to that of the imperfect and perfect (aorist perfect) indicative, a common misconception (shared by Morwood, LG, p. 99, to judge from the examples he gives and his remark ‘the tense is dictated by the sense’. See also Taylor at LB, p.16). The perfect seems to be used where Greek uses the indicative (Gildersleeve & Lodge §513 NOTES -1 seem to think it replaces the Greek infinitive). Whether the imperfect is used to suggest the same kind of result as the Greek infinitive is unclear, since what kind of result the infinitive represents is unclear. Woodcock (surely still the best guide in English to actual Latin usage) says (A New Latin Syntax, p. 122) that the imperfect is used to show the ‘logical connexion’ between a cause (stated in the main clause) and an effect (stated in the result clause), which suggests that the effect/result is treated as a necessary one, while the use of the perfect subjunctive denotes a merely contingent effect/result.Footnote ⁷ If this is really what Roman writers supposed that they were doing, then both they and Woodcock were mistaken. The truth is that there is no logical connection between facts, only between propositions and the terms of propositions. Also, there are no such facts as necessary facts (though there are necessary truths), and, as David Hume showed, there is no necessary connection between a cause and its effect. Ironically, if Woodcock were right and the imperfect denoted a necessary result, it would surely have been used of an actual result, which is what the perfect subjunctive is supposed to represent. For in what sense can an effect be necessary but not actual? I conclude that we still don't know exactly what kind of result the Latin imperfect subjunctive (or the Greek infinitive) represents.

(Note that the difference, whatever it is, marked in Latin by the use of the perfect and imperfect subjunctive, refers to results in past time only, unlike the Greek use of the infinitive, even though a result clause can be used of present and future time. Apart from the fact that a result in the future cannot be an actual one from the perspective of the present, this is a consequence of the Latin choice to mark the difference by tense rather than mood: there is only one present subjunctive for present time.)

The use of the Latin perfect subjunctive tends to be thought of in the same way as the Greek indicative, i.e. as a guarantee of the actual occurrence of an action or event rather than as an expression of the writer's attitude to the facts. Similarly with the imperfect subjunctive (on the part of those who are aware that it is not generally used with the same sense as the imperfect indicative), we can know what kind of result it is from the way in which the writer expresses himself. There is disagreement, however, on what kind of result this is — as well there might.

Latin Causal Clauses

‘A causal clause indicates the reason why something happens. The logic of the indicative is that it states a factual reason: there is no doubt about the cause and effect relationship …’ (Taylor, LB, p.36) (Taylor does not say whether the absence of doubt is that of the writer or reader, but in either case he is not in a position to say this.)

‘… to contrast a rejected reason with a true one’ (p.36)

‘A causal clause states a reason why something happens’ (p. 37) (If ‘cause’ and ‘reason’ are interchangeable, does this mean anything more than that a causal clause states a cause?)

As with Greek, Latin distinguishes a cause given by the writer from a cause attributed by the writer to someone else. As usual, Woodcock's account (A New Latin Syntax, p. 196) is the clearest and fullest in English — but even Woodcock confounds ‘cause’ and ‘reason’. The difference between Greek and Latin is that in Greek the distinction made by the indicative and the optative can only be made if the main verb is in a historic tense. In Latin the distinction made by the indicative and subjunctive can be made if the main verb is in a historic or primary tense.

There is the same tendency with the Latin construction as with the Greek: to treat the cause given by the writer in the indicative as an actual cause as opposed to an ‘alleged’ (erroneously) cause in the subjunctive.

Latin Comparative Clauses

Comparative clauses may be factual (a loaded term!) with a verb in the indicative or imaginary (another loaded term!) with a verb in the subjunctive. Other terms than ‘factual’ and ‘imaginary’ may be found to denote the difference. (See below for the difference between a factual and an imaginary comparison.) Also, the tense of the subjunctive used after a primary main verb differs depending on the writer's attitude to the facts, in this case whether the writer wishes to suggest that the situation envisioned in the comparative clause is likely to be the case or not. This second distinction is not often come across in grammar and course books, or is not well explained if it is. One result of this is that users of such books encounter instances of verbs which appear to be in the ‘wrong’ tense according to the (only) rule they have been given. The sequence of tenses rule they are usually given fails to account for the actual usage of Latin writers.Footnote ⁸

So, if the writer wishes to suggest that an imaginary situation X was not/is not/will not be the case, he expresses this in the form of the protasis of a past unreal condition, present unreal condition and periphrastic ‘future in the past’ subjunctive respectively, preceded by a comparative word such as tamquam, velut or one of the myriad of other comparative expressions. (The periphrasis is used instead of the form of the protasis of an ideal condition because the latter (present subjunctive) would indicate present not future time (under the sequence of tense rule used to suggest a likely situation), and would not convey the suggestion of unlikelihood.) Because the comparative clauses have the form of unreal conditions, with the same (assumed) implication that X was not or is not the case, there is a tendency to assume that the opposite of what is said in the comparative clause was or is a fact, as with actual unreal conditions in Greek (and Latin too). And, of course, as we shall see, particularly from the extract from Morwood that follows, factual comparisons using the indicative are taken to be facts rather than the writer's expression of his attitude to the facts, an attitude that may be mistaken and known by the writer to be mistaken.

Morwood's (attempted) explanation (LG, p. 128) of the distinction between a factual and an imaginary comparative clause is probably the best example I have come across of the confounding of a fact with the statement of a fact.Footnote ⁹ It needs to be quoted in full (the italics are mine):

‘The senators were terribly afraid, as if the enemy were already at the gates of Rome.’

‘The general was rewarded as his courage deserved.’

In the first of these sentences, the comparison is untrue. The enemy were not at the gates of Rome. In the second sentence, the comparison is true. The general's courage did deserve to be rewarded.

In Latin if the verb conveys a fact (as in the second meaning above), it is naturally in the indicative, since it is true. If the verb makes an imaginary (i.e. untrue) comparison (as with the first meaning above), it is in the subjunctive. Comparisons are much more likely to be untrue than true.’ [end of quotation]

The writer (and the senators) might have been unaware that the enemy were already at the gates of Rome. As for the second sentence, the writer might have been mistaken on several counts.

Factual comparisons compare what is stated in the main clause with an alleged fact stated in the subordinate clause of comparison. The alleged fact may not be a fact at all. The comparison may be false. The writer may know that it is false. Even the main clause may be false. Imaginary comparisons compare what is stated in the main clause with an imaginary situation hypothesised by the writer, either implied to be unlikely or with no such implication, especially if the main verb is in a primary tense, as explained above. Again the main clause may be false.

Often an imaginary comparison may not only be true (pace Morwood et al.), the speaker may actually imply that it is true. The sentence ‘You're talking as if you know something about it’ may, in a certain context, imply that the addressee does know something about it — and in fact he does. (It may of course, in context, imply the opposite — and be mistaken.) In certain situations Latin can resolve, or at least reduce, the ambiguity of English by means of the tense of the subjunctive used: scires in our example above would imply the addressee was ignorant; scias would carry no such implication. What Latin cannot do, what no language can do, except in the case of certain verbal paradoxes, is to self-certify its truth or falsehood by means of the forms it assumes.

Latin Concessive Clauses

Concessive clauses, like comparative clauses, may be factual or imaginary. But the basis of the indicative/subjunctive distinction does not apply to concessive clauses, as factual concessions have their verbs in the subjunctive (in the appropriate tense) after quamvis, licet, cum and qui. (quamvis may also be used to introduce imaginary concessions, and is more often used in this way.)

Imaginary concessive clauses introduced by compounds of si, i.e. etsi, etiamsi, tametsi, have the forms of the protases of unreal and ideal conditions. This is not the case with quamvis (see Taylor's comment in LB, p.37). Factual concessions introduced by compounds of si have the forms of the protases of open conditions, with verbs in the indicative in the appropriate tense, like quamquam.

Concessions that have the same forms as the protases of unreal conditions are liable to the same misinterpretations as unreal conditions (q.v.): an implication is taken for a statement and a statement for a fact. Concessions that have the same forms as the protases of open and ideal conditions are not likely to be misinterpreted in this way.

Concessions may also be expressed in the form of consecutive sentences (ita … ut/ut non … or ut/ut non … tamen …), comparative sentences (ut … ita …) and the simple iussive subjunctive. None of these is likely to be misunderstood in this way either.

Latin Conditional Clauses

‘The things supposed may be either ones which did not happen, are not happening now …’ (Morwood, Writing Latin, p. 91)

‘A past closed condition says if X had happened, Y would have happened [Taylor's italics] (but in fact neither did)’ [my italics]

‘A present closed condition says if X were happening now, Y would be happening [Taylor's italics] (but in fact neither is)’ [my italics] (LB, p. 48)

‘The notion of impossibility comes from the irreversible character of the Past Tense.’ (Gildersleeve & Lodge §597) What is impossible to reverse, of course, is not a feature of a verb but an action or event. Even allowing for hypallage, there is a conflation here of language with what language represents: as we have seen, a common slippage between one and the other.

Much the same applies to conditional clauses in Latin as in Greek. We find the same types of conditions in Latin as in Greek — what I call ‘Open’, ‘Unreal’ and ‘Ideal’. Again as with Greek, it is with the Unreal conditions that there appears to be misunderstanding, and for the same reasons. To repeat, assuming that the writer is implying (he is not stating) that X and Y did not happen/are not happening, he may be mistaken in his implication in either case (perhaps he is not in a position to know) or he may be lying. If he is, then it is not a fact that X and Y did not happen/are not happening.