2.1.1 Methods of Analysis

Before we begin, I should explain certain issues of methodology. I use a range of analytical methods in this chapter. Discussions of individual examples serve to illustrate theoretical points. In other cases, collections of examples, or contrastive case studies, are used to suggest a pattern where the underlying variables are hard to quantify. Where possible, however, I try to corroborate my argument with quantitative analyses. With respect to these analyses, three main issues must be addressed here so that the reader may form a clear understanding of the meaning of the presented results. First is the matter of the selection of text passages (Section 2.1.1.1); second, the trimming of the data to ensure a focused comparison (Section 2.1.1.2); and third, the interpretation of the relevant statistics (Section 2.1.1.3).

2.1.1.3 Interpreting the Statistics

The second main issue with previous quantitative analyses of the present for preterite in Classical Greek is that they do not provide the statistics required to adequately assess the potential meaning of a distribution. For example, Allan (Reference Allan and Lallot2011a: 40) reports that 51 of the 150 clauses containing a present for preterite in his corpus were introduced by the particle καί (‘and’), while this was true only for 33 of the 150 verbs in the contrastive corpus. The meaning of this distribution depends upon whether the contrast is made with another specific particle or with the ‘rest’ category; for the present discussion, I will assume a contrast with the particle δέ, another form of ‘and-coordination’ (Bonifazi et al. [Reference Bonifazi, Drummen and de Kreij2016: IV.2, §1]; for the contrast between καί and δέ, see Allan [Reference Allan and Lallot2011a: 41–2]).

If we compare the figures for καί with those for δέ, we find the following distribution:

This distribution suggests that the odds of the present being used are stronger when the particle καί introduces the sentence than in the case of δέ. What is missing at this point is a measure of the size of the effect. Moreover, we have no indication of how likely it is that this distribution is merely the result of chance.

In my analyses, I use the odds ratio to express the effect of a variable on the tendency for the present to be used instead of the preterite. The odds ratio is calculated as follows. First, we take the odds of the present being used under a baseline condition. What the baseline condition is depends on how we formulate the hypothesis. If we hypothesise that the particle καί as discourse connector increases the odds of the present being used in comparison with δέ, we take δέ as the baseline condition.

Under the baseline condition, the odds of the present being used are 54/60 = 0.9. Then we calculate the corresponding odds under the ‘experimental’ condition. Here the odds are 51/33 = 1.545. The odds ratio is the ratio between these odds, that is, 1.545/0.9 = 1.717. This means that the odds of the present being used instead of the other tenses increase by a factor 1.717 when καί is used as a discourse connector instead of δέ. (This value is not affected by the fact that, in reality, the ratio between the total number of present forms against the total number of other tense forms is much different from that in the sample used for this analysis.)

Whenever we look at a certain distribution in a limited dataset, there will always be some variability due to chance. A distribution is only statistically significant when the probability of that distribution occurring simply by chance is very low. This ‘p-value’ can be calculated by means of logistic regression (I used IBM SPSS statistics to run this procedure). In the present case, the p-value is 0.064.Footnote ²⁶ That means that there is a chance of, roughly, 1 in 16 to obtain a distribution such as the one observed here by chance. It is still relatively uncertain, then, that the observed distribution is meaningful – the value of 0.064 is above the (arbitrary) rule-of-thumb value of 0.05 for statistical significance.

While I report p-values in my analyses, it should be clearly understood that these do not have the same status as they would in a proper experimental study. It is impossible to generate new data for Classical Greek. The texts on which I test my hypotheses are the same texts that have partly informed these hypotheses, either through the work of other scholars or through my own observations. The reason this is problematic lies in the nature of significance testing. A significance test tells us how likely it is that a certain distribution is the result of chance alone. However, if we test multiple related hypotheses on the same data, we are more likely to find an unusual distribution that is merely due to chance. It is like rolling a hundred-sided die: the chance of rolling a particular number is 1 in 100, but if we roll the die a hundred times, this chance increases dramatically.

This is why proper hypothesis testing requires the hypothesis to be formulated before looking at the dataset on which it is tested. If we mine a certain dataset for interesting distributions, we are essentially ‘rolling the die’ an unspecified number of times. When we then do a significance test on this distribution in the same dataset, it remains wholly obscure how many hypotheses were implicitly rejected (because it was found that those distributions were not interesting). It is therefore unclear how meaningful the resulting p-value is.

To justify and explain my reporting of p-values here, let me first note that I did not perform ‘data dredging’. This is the procedure of automatically testing an exhaustive number of possible correlations in a certain dataset (‘rolling the die’ thousands of times). I only explored a limited number of possibilities, and my hypotheses were largely informed by the work of other scholars, some working on different languages (e.g., Kroon [Reference Kroon, Sawicki and Shalev2002] on Latin) or on different parts of the Classical Greek corpus (e.g., Allan [Reference Adams2007], [2011a] on Thucydides). Moreover, the hypotheses form a theoretically coherent whole. The point is that I did not arbitrarily go looking for any correlations that might turn up in the data.

Now, the fact remains that the p-values cannot be properly corrected for the total number of tests I ran, so that they only represent the intrinsic probability associated with the reported distributions. I believe that this is justifiable given the present conditions (the impossibility of generating new data), so long as it is properly understood what the p-value represents. The point of reporting p-values is purely to give some indication of the saliency of the distributions reported here. If I would just report the counts (so many present forms against so many aorists under this or that condition), the reader would have no idea whatsoever whether such a distribution might actually mean something.

More specifically, the function of p-values in the present context is threefold. First, it guards us against drawing conclusions too quickly. If the p-value is high even under the present circumstances, then this means that we should be extra skeptical of attaching any value to a certain distribution.

Second, it is useful to compare different p-values. No matter how many tests we run on the same dataset, a difference between p-values is always instructive. If one test yields a p-value of 0.01, while another yields a p-value of 0.001, we know at least that the latter distribution is more likely to be meaningful than the former.

Third, I would maintain that even in the present context, a very low p-value is strongly suggestive of a meaningful correlation, granted that the hypothesis in question is simple and clearly motivated, and the counts yield a substantial odds ratio.

2.2.1 Verb Type

If the concreteness of the conceptualised event enhances the strength of the simulation, and if present for preterite usage is correlated to simulation strength, then we should expect the odds of the present for preterite being used to increase when the verb phrase designates a concrete action.

To test this hypothesis, one would, ideally, compile a list of all verbs occurring in dramatic narrative and rate them on a scale of concreteness and then proceed to look for a correlation between scores on this scale and tense use. In practice, this procedure is unworkable, or at least requires a large-scale effort that transcends the bounds of this study. There are just too many verbs, and determining points on a concreteness scale for verbs is a problematic procedure – more so as verbs have different meanings and metaphorical usages.

A survey of the corpus does show, however, that most verbs that have a high present-to-aorist ratio (compared to the general baseline of 0.43) designate highly concrete events. I present the most salient findings in Table 2.1. Only verb groups that yielded at least 4 instances were included. There is some degree of arbitrariness in the groupings, but the rationale behind them will, I hope, be readily apparent (for example, I subsume ‘hit’, ‘strike’ and ‘stab’ under a single heading).Footnote ³¹ Note that the present forms of this select group of verbs amount to almost two-fifths of the total number of present forms in the corpus (72/183 = 0.39).Footnote ³²

Table 2.1Tense and verb types

^a I have not included cases where the verb is used intransitively (S. El. 902 ἐμπαίει [‘hits’], OT 1252 εἰσέπαισεν [‘crashed into’]).

Let me discuss these findings in more detail, breaking my discussion down in four groups: transitive action, intransitive movement, perception and sound production.

1. (a) Transitive action. Five of the verb groups listed in Table 2.1 designate transitive action: ‘hit’, said of a projectile or a person sending forth a projectile; ‘hit/strike/stab’, with the hands or with an instrument; ‘throw’; ‘loosen’; and ‘draw/pull’. All events are highly concrete, and except for ‘loosen’, all events involve strong physical exertion.

2. (b) Intransitive movement. The verb πίπτω (‘fall’) is used in two ways. It can refer to unintentional movement, as in Euripides, Phoenician women 1186 ἐς γῆν δ’ ἔμπυρος πίπτει νεκρός (‘and the flaming corpse falls to the ground’). When the verb refers to intentional movement, it may be translated as ‘throw oneself (at)’, ‘rush (at)’, or ‘crash (into)’: for example, Euripides, Medea 1205 πατὴρ δ’ ὁ τλήμων … προσπίτνει νεκρῷ (‘and the miserable father throws himself at the corpse’). In both cases, the verb designates quick, intense movement, and in the case of unintentional movement, violence is often involved (someone falls after being hit).Footnote ³³

3. (c) Perception. The ‘see’ group yields a high present-to-aorist ratio (2.33). For ‘hear’, the ratio is 1; however, 2 out of the 3 aorists are used in a peculiar way. After a character has made a speech, the aorist signals that (an)other character(s) heard it: see, for example, Euripides, Phoenissae 1437–9 στέρνων δ’ ἄπο | φύσημ’ ἀνεὶς †δύστλητον† Ἐτεοκλῆς ἄναξ | ἤκουσε μητρός (‘letting out a gasping breath from his chest, lord Eteocles heard his mother’).Footnote ³⁴ What matters here is the mere fact that Eteocles actually heard the speech. Compare an instance such as Euripides, Rhesus 789 κλύω δ’ ἐπάρας κρᾶτα μυχθισμὸν νεκρῶν (‘lifting my head, I hear the groaning of men dying’). Here the subject is suddenly confronted with a striking sound. Taking this into account, the present-to-aorist ratio of the ‘hear’ group will be closer to that of the ‘see’ group. In both cases, the predilection for the present can be understood in terms of the concreteness of simulated perception.

4. (d) Sound production. The frequent use of the present for preterite of the verb βοάω (‘shout’) seems unsurprising ‘because of its inherent dramatic character’ (Allan [Reference Allan, Bakker and Wakker2009: 192]). In terms of concreteness, shouting involves physical exertion (for the agent) and the sound that is produced strongly imposes itself on the senses (of the bystanders).

However, many other verbs designating sound production are never marked with the present for preterite in the corpus. This is true for verbs formed after interjections, verbs that designate less specific noise made by humans or animals and verbs designating noise in general. Examples are οἰμώζω (‘cry oimoi’), ἀλαλάζω (‘cry alalai’), ὀτοτύζω (‘cry ototoi’), εὐφημέω (‘cry euphemei’, ‘cry out’), κράζω (‘scream’), κλάζω (‘scream’), στενάζω (‘groan’), θορυβέω (‘make noise’), ῥοθέω (‘make a rushing noise’), κτυπέω (‘resound’) and expressions meaning ‘clap the hands’ (κροτέω/κρούω χεῖρας).

Does this falsify the idea that the predilection of the verb βοάω (‘shout’) for the present for preterite is due to its dramatic character? Or is the threshold for present tense usage raised for these other verbs by extraneous factors? In my view, the latter explanation is more likely. To begin with, in the one instance where the present is used with a verb similar to the ones listed, the intensity of the sound seems a crucial factor:

Antigone cries out ‘bitterly’ (πικρῶς), producing a ‘piercing sound’ (ὀξὺν φθόγγον).Footnote ³⁵ It is probably not a coincidence that the present is used precisely where the character of the sound is particularly salient.Footnote ³⁶

I believe the crucial difference between βοάω (‘shout’) and most of the verbs listed above lies in actionality. Verb phrases of ‘making noise’ are typically of the activity type, which means they designate events that can be extended indefinitely. Clear examples are θορυβέω (‘make noise’), ῥοθέω (‘make a rushing noise’). The same is true for verb phrases meaning ‘clap the hands’ (κροτέω/κρούω χεῖρας) and verbs modelled on interjections: cries such as ὀτοτοῖ could be extended and repeated ad libitum (e.g., ὀτοτοτοτοῖ, ὀτοτοῖ ὀτοτοῖ; see the entry in LSJ). The use of the present for preterite is dispreferred with activity verb phrases in the corpus.Footnote ³⁷ The verb βοάω (‘shout’), by contrast, is probably more easily understood to designate a single shout. Still, I must admit I do not have a particular explanation for why we do not find the present for preterite with the verbs κράζω (‘scream’) and κλάζω (‘scream’) in the corpus (8 instances).Footnote ³⁸

All in all, I believe the data presented here supports the idea that the odds of the present for preterite being used are increased when the designated event is concrete (concreteness being understood in terms of the involvement of sensorimotor processes) and intense (in terms of the degree of exertion involved in a physical action or the strength of the impression made by a sight or sound).

2.2.2 Reference to the Limbs

If an action is explicitly said to involve the use of the hand/arm or foot/leg, then the underlying conceptualisation will be highly concrete. According to the argument made in this chapter, this should raise the odds of the present for preterite being used. This hypothesis was initially based on incidental observations. It then turned out that there were only a few instances in total. I have looked at instances of the words χείρ (‘hand’, sometimes ‘arm’ by extension), πούς (‘foot’), κῶλον (‘leg’), σκέλος (‘leg’), γόνυ (‘knee’). The rule was that the noun had to be connected to the main verb, either as subject, or as instrumental dative (‘with the hand’) or as accusative of direct object.

For the word χείρ (‘hand’), the instrumental dative yields 5 instances of the present against 4 of the aorist.Footnote ³⁹ There are some interesting contrastive pairs. Compare Euripides, Hippolytus 1188 μάρπτει δὲ χερσὶν ἡνίας ἀπ’ ἄντυγος (‘he grabs the reins from the rim with his hands’) with Phaethon 172 (Diggle [1970]) τοσαῦτ’ ἀκούσας παῖς ἔμαρψεν ἡνίας (‘having heard that much, his son grabbed the reins’).Footnote ⁴⁰ And again, Sophocles, Women of Trachis 923–4 τοσαῦτα φωνήσασα συντόνῳ χερὶ | λύει τὸν αὑτῆς πέπλον (‘having said that much, she loosens her robe with a swift hand’) versus Oedipus at Colonus 1597 εἶτ’ ἔλυσε δυσπινεῖς στολάς (‘subsequently, he loosened his dirty clothes’). The direct object gives us 1 present and 4 aorists.Footnote ⁴¹ This amounts to a total of 6 presents against 8 aorists; in three instances of the latter group, however, the verb phrase is of the activity type, where the present is dispreferred.Footnote ⁴²

I have found only 6 relevant references to the foot, leg, or knee: 2 present forms and 4 aorists forms.Footnote ⁴³ A comparison of the use of the present with that of the aorist will serve to suggest that concreteness is the decisive factor. The two instances of the present are found in the narrative of the second messenger in Euripides’ Phoenician women, which tells of the duel between Eteocles and Polynices. First, Eteocles inadvertently exposes his leg:

The present τίθησι (‘places’) designates a concrete movement of the leg. Note also the mention of the foot (ποδί [‘with his foot’]) in the participial clause. The other instance occurs further on, where Eteocles makes a special move:

Again, the movement designated with the present ἀναφέρει (‘brings’) is highly specific.Footnote ⁴⁴

Now let us contrast the aorists. In Euripides, Suppliant Women 718, the phrase is ἔτρεψαν ἐς φυγὴν πόδα (‘turned their foot to flight’), which is a poetic periphrasis for ‘fled’ (compare Section 2.4 on narrative mimesis and grammatical simplicity). No specific motion of the foot is intended. In Iphigeneia amongst the Taurians 1407, we read χὠ μέν τις ἐς θάλασσαν ὡρμήθη ποσίν (‘and one rushed to the sea with his feet’). Again, this does not refer to a specific motion. In lines 332–3 of the same play we read ἐς δὲ γῆν γόνυ | καμάτῳ καθεῖσαν (literally ‘they sent down their knee to the earth in exhaustion’). This is another poetic periphrasis.Footnote ⁴⁵ We may contrast the use of the present in Hecuba 1150, where the knee is mentioned in a subordinate clause: ἵζω δὲ κλίνης ἐν μέσῳ κάμψας γόνυ (‘I sit down in the middle of the couch, bending my knee’).

In conclusion, a review of instances where the limbs are mentioned supports the argument that the present for preterite is more likely to be used when the designated event is strongly concrete.

2.2.3 First Person versus Third Person

Comparing tense usage with the first and third person in the corpus of dramatic narrative is problematic because of morphological ambiguity. As I explained in Section 2.1.1.2, I have cut from the data forms that are morphologically ambiguous between the present and the imperfect (see also Introduction, Section I.4.3). Most of these forms are in the third person (about thirty; see Appendix, Section A.2.2.1). This means that, in the current dataset, the distribution will be skewed so that it will seem that the present is used less often in the third person than it actually is.

There is another aspect to this issue. For a number of individual instances, even if we assume they are present forms, the implicit aspectual construal may still be imperfective. In the following example, we can be relatively certain of this. Heracles, in his madness, has just killed two of his children, and now moves against the third:

The form χωρεῖ (‘moves’), printed by Diggle with the accentuation of the present tense, is actually morphologically ambiguous: it might also be an imperfect without the augment, (ἐ)χώρει (‘moved’). I think Diggle’s reading is probably right here (see Section 2.6.2). Even so, the implicit aspectual construal here is most likely imperfective. The imperfective is typically used with verbs of movement when no destination is specified. With χωρέω (‘move’), this is true even when a destination is specified – in fact, I only found one aorist of this verb in the corpus.Footnote ⁴⁶

It is hard to make a confident estimate of the number of present forms with perfective aspectual construal in this sample. In at least seven cases, I believe we should probably read the imperfect.Footnote ⁴⁷ Nine of the remaining instances are of the verb χωρέω (‘move’) which, as I just argued, has a predilection for imperfective construal. All in all, I think an estimate of fifteen aoristic present forms among the thirty ambiguous instances of the third person is reasonably generous.

On this assumption, the figures and statistics are listed in Table 2.2 (the numbers between brackets represent the expected counts, that is, the numbers we would expect to find if the distribution were completely even). As we can see, the actual number of present forms with first person marking exceeds the number expected under an even distribution; the converse is true for the aorist. The size of the effect is expressed by the odds ratio: the odds of the present being used increase by a factor 1.916 when the verb is in the first person. Finally, the p-value represents the probability of finding this distribution by chance. (It should at all times be remembered that this value is not corrected for the total number of tests; see Section 2.1.1.3.) As the odds ratio is substantial and the p-value is relatively low, these figures suggest that there is a correlation between present tense usage and first person marking.

Table 2.2Tense and person marking

^a B = .65 (constant = -.893), standard error = .211.

I should note that the connection between first person marking and memory is problematic due to the nature of the texts studied here. It is the playwright who writes the text, but the memory belongs to the character; moreover, the actual speaker of the words, the actor, has no actual memory of the narrated events. The hypothesis concerning tense and person marking put forward here is based on the assumption that art imitates life and that the language sense of the playwright will make his characters’ speech like actual speech. If this premise is accepted, the results presented here lend support to the idea that vividness of memory positively correlates with the use of the present for preterite (compare Park et al. [Reference Park, St-Laurent, McAndrews and Moscovitch2011]).

2.2.4 Conclusion

I have argued in this section that the use of the present for preterite is positively correlated with narrative mimesis in terms of mental simulation: the higher the activation strength of the simulation, the more the conceptualisation will feel like actual experience and the more the conceptual distinction between the actual past events and the present simulation will be blurred.

This idea produced two specific hypotheses. First is the idea that the concreteness (involvement of actual sensorimotor processes) and intensity (involvement of strong physical effort or striking audio-visual components) of the designated event increase the odds of the present for preterite being used. I have discussed tense usage with certain verb types (Section 2.2.1) and in descriptions that make explicit reference to the limbs (Section 2.2.2) to corroborate this claim. Second, I suggested that the simulation will be stronger when the speaker was actively involved in the described events, and thus the odds of the present for preterite being used in these cases will be increased. This idea is supported by the corpus data (Section 2.2.3).

2.3.1 Speech-Reporting Verbs

Let us now consider the data from the corpus with respect to speech-reporting verbs. To begin, I hypothesise that verbs introducing direct speech have higher odds of being marked with the present than those that do not.Footnote ⁵⁵

The results, listed in Table 2.3, do not support the hypothesis. The odds of the present being used are in fact lower with verbs introducing direct speech than with verbs that do not; however, the effect is relatively small (odds ratio 0.716) and statistically highly uncertain (p = 0.213).

Table 2.3Tense and direct speech representation

^a B = -.334 (constant = -.796), standard error = .268.

In retrospect, a problem with this hypothesis is that the verbs in the baseline condition are entirely heterogeneous: some have a strong predilection for the present for preterite, while for others, the opposite is true (Section 2.2.1). It may be that the odds of the present being used in the case of verbs introducing direct speech is relatively high with respect to some verb groups but low with respect to others. Still, we might have expected verbs introducing direct speech to rank very high on this scale due to the strongly mimetic quality of direct quotation, but this is not what we find.

Whether the present for preterite is used seems to be determined by the character of the speech and its function in the context. Most typically, the present marks the speech of characters who are indignant or distressed. An example will serve to illustrate the point. In Sophocles’ Philoctetes, Neoptolemus tells Philoctetes a made-up story of what happened when he arrived at the Greek camp in Troy. Agamemnon and Menelaus told Neoptolemus that the weapons of his father Achilles had been given to Odysseus (363 εἶπον [‘said’]). Neoptolemus reacts with indignation (368 λέγω [‘say’]). Odysseus responds that the Atreids were right in their decision (371 εἶπε [‘said’]). When Neoptolemus starts abusing the leaders (374–6, no direct speech), Odysseus responds with a jab of his own (378 ἠμείψατο [‘answered’]). Only Neoptolemus’ first speech is marked with the present, and a look at the actual text makes clear why:

The narrative here has a markedly mimetic character, which stands in contrast to the other speech introductions. The speech introduction is preceded by a main clause denoting sudden (εὐθύς [‘immediately’]) and concrete bodily movement (ἐξανίσταμαι [‘stand up from my seat’]). Moreover, Neoptolemus emphasises his emotional distress at this point: he ‘broke out in tears’ (ἐκδακρύσας) and felt ‘heavy anger’ (ὀργῇ βαρείᾳ) and ‘pain’ (καταλγήσας). Finally, the fact that the subject is the speaker is also an important factor (Section 2.2.3).Footnote ⁵⁶

A more focused contrast is that between verbs introducing direct speech with those introducing indirect speech: this zooms in purely on the mimetic qualities of the reported speech.Footnote ⁵⁷ Here we are confronted with some difficulties. First, with indirect speech, there are 0 present forms against 9 aorist forms.Footnote ⁵⁸ While this seems telling in itself, a count of zero makes it impossible to calculate a meaningful odds ratio and p-value in logistic regression.

Also, there is the problem of morphological ambiguity (compare Section 2.2.1). A number of verbs that are ambiguous between the present and the imperfect in the third person singular happen to belong to verbs that introduce speech. Five of these introduce direct speech; for three of these it is reasonable, in my view, to suppose that the tense is in fact present, based on present tense usage in the immediate context. An example is found in the messenger narrative in Sophocles’ Antigone, where the messenger tells how Antigone was found dead and Haemon killed himself. As Creon enters the tomb, he hears a terrible shout, and reacts by crying out (1210–11): οἰμώξας δ’ ἔπος | ἵησι δυσθρήνητον (‘wailing, he sends forth words full of bitter lament’). This is followed by direct speech. The present form here marks the speech of a distressed character (see above) and is the fourth in a series.Footnote ⁵⁹ When Creon enters the tomb, seeing his son embracing a dead Antigone, he acts as follows:

To my mind, there need be little doubt that both χωρεῖ (‘moves’) and καλεῖ (‘calls out’), both of which are morphologically ambiguous, are in fact present forms. This fits the intensity of the overall narrative, and with respect to the latter form, the parallel with the present ἵησι (‘sends forth’) in line 1211 is strong.Footnote ⁶⁰

One ambiguous form may be considered to introduce indirect speech, although it is problematic in several respects.Footnote ⁶¹ We may include it in the data for the sake of removing the 0 count for present tense and indirect speech, so that we may run a regression analysis. Based on these premises, the results are as listed in Table 2.4.

Table 2.4Tense and speech representation

^a B = 1.201 (constant = -2.197), standard error = 1.081.

The effect is large (odds ratio 3.323) but statistically barely meaningful (p = 0.267). The obvious problem is that there are so few cases of indirect speech representation in the corpus.

2.3.2 Conclusion

Of the four modes of depiction discussed in this section (gesture, stress and intonation, sound symbolism and direct speech representation), only direct speech representation was relevant to the present investigation. There is no evidence that direct speech representation increases the odds of the present for preterite being used with the reporting verb in comparison to other verbs. Comparing direct speech representation to indirect speech representation does yield an effect, but this is statistically highly uncertain. However, certain specific examples suggest that mimesis in terms of the character of the reported speech (mimicry: [10]; intensity: [12], [13]) does play a role in present for preterite usage.

I close this section by noting some peculiar tendencies pertaining to tense usage and certain speech-introducing formulae outside the selected corpus. The first type is found in Herodotus’ Histories (compare Eriksson [Reference Eriksson1943: 12]). In a number of dialogue passages, Herodotus uses speech introductions without a connective particle (‘asyndeton’), which is uncommon in Classical Greek. In these formulae, the present is the tense of choice. To take the case of the verb ἀμείβεσθαι (‘answer’), we find 12 instances of the present with asyndeton against 1 imperfect (7.16.2; no relevant aorist forms). The verb always stands at the beginning of the clause, for example: ἀμείβεται Κροῖσος τοῖσδε (‘Croesus answers with the following [words]’; 1.38.1).Footnote ⁶² Sometimes, there is only the verb of speaking and the name of the speaker: ἀμείβεται Κροῖσος (‘Croesus answers’; 1.40.1).Footnote ⁶³ Finally, there may be anaphoric reference to the words of the previous speaker: ἀμείβεται πρὸς ταῦτα Δαρεῖος (‘to that Darius answers’; 3.140.4).Footnote ⁶⁴ In my view, the absence of a connective particle here is a sign of the narrator trying to minimalise his presence and make the characters ‘speak for themselves’ as much as possible (see Section 2.4 on ‘mimetic’ grammar). Moreover, this use tends to occur in highly salient dialogues, such as Croesus’ tragic dialogue with his son (1.38.1, 1.39.1, 1.40.1) or the so-called constitutional debate by the Persians who are planning to overthrow the government (3.134.6, 3.140.4). These are fitting contexts for a mimetic, involving representation of what was said (compare Section 2.5.1 on communicative dynamism).

My second observation concerns the verb φημί (‘say’), the standard verb in inquit-formulae (i.e., where the reporting verb follows words of the speech representation: ‘What’, said he, ‘do you think?’). In the plays of Menander, the use of the present instead of the preterite is almost a rule with this verb in narrative.Footnote ⁶⁵ I reproduce part of example (1) cited in Section 2.1 to illustrate this use:

The practice of Aristophanes mirrors that of Menander: I have found three instances of the present for preterite of φημί (‘say’) in inquit-formulae, and 0 of the imperfect.Footnote ⁶⁶ We do not find this in other genres: in tragedy, the verb is hardly used at all, and never with the present for preterite; in historiography, the verb is common, but it is never used with the present for preterite; and in Attic rhetoric, the verb is again common, but there are only three instances with the present for preterite (all in inquit-formulae).Footnote ⁶⁷ This suggests that the phenomenon is a colloquialism and that tense-switching may be more sensitive to direct speech representation in natural than in artificial narrative.

2.4.1 Discourse Time and Story Time

The relation between discourse time and story time is hard to quantify because both variables are elusive.Footnote ⁸⁰ The time it takes for something to be read or spoken is not fixed. Story time can sometimes be determined with certainty, when the time it took for something to occur is made explicit, or when the event is punctual (e.g., ‘sneeze’). Most of the time, however, the exact duration of the designated events is uncertain. Given these difficulties, I will not present a quantitative analysis but only discuss some illustrative passages.

When the narrated events are punctual and follow closely upon one another (typically with a causal relationship between them), the present for preterite seems to be preferred (compare von Fritz [1949: 186–7] on a passage in the Roman poet Catullus). This is illustrated by example (3), which I reproduce here:

The moment the speaker’s head emerges outside the city walls (προκύπτω [‘stoop forward’]), he is struck (παίει [‘hits’]); this immediately causes him to fall (πίπτω [‘fall’]). We may assume that the speaker’s intention to shout (μέλλω τε βοᾶν [‘I am about to shout’]) follows immediately.

A similar case is the following (compare [17]):

As Eurydice is loosening the bolts of the gate, she is struck by a sound (βάλλει [‘strikes’]). This causes her to fall into the arms of her slaves (κλίνομαι [‘recline’]) and faint (ἀποπλήσσομαι [‘faint’]). Again, the present forms designate punctual events occurring in immediate succession.Footnote ⁸¹

I have not found clusters of aorist forms in similar passages where punctual events follow so immediately after one another. The following passage illustrates how the aorist is used when events evolve at a more leisurely pace:

In (18) and (19), one event immediately causes the other’s occurrence, while the actions described in (20) merely constitute steps in a protocol. (Narrative tempo is also related to communicative dynamism [Section 2.5.1]: events tend to evolve quickly at critical points in the story.)

2.4.2 Discourse Connection

I make a distinction between retrospective discourse connection, where particles specify the relation between the host clause and the previous one, and prospective discourse connection, where particles anticipate a relation between the host clause and the following one.

For present purposes, three modes of retrospective discourse connection can be distinguished: simple ‘and-connection’; complex discourse connection (particles designating adversative or hierarchical relationships); and no explicit connection (‘asyndeton’). I formulate three hypotheses with respect to these types of discourse connection.

First, particles specifying complex discourse relations are expected to lower the odds of the present for preterite being used in comparison with simple ‘and-connection’. Particles belonging to the former group are the following: γάρ (‘for’), which has an embedding function (a ‘push’ particle); οὖν/(τοί)νυν (‘so’), which brings the discourse back to a higher level (a ‘pop’ particle); and ἀλλά (‘but’) and μέντοι (‘however’), which signal adversative relations (see Slings [Reference Slings and Rijksbaron1997]; Bakker [2011]). The particles δέ, καί and τε constitute the baseline group of ‘and-connectives’. The results of the comparison are listed in Table 2.5.Footnote ⁸²

Table 2.5Tense and simple versus complex discourse connection

^a B = -.269 (constant = -.795), standard error = .379.

The observed effect (odds ratio 0.764) is statistically meaningless (p = 0.477).

Second, within the group of and-connectors (δέ, καί and τε), the particle δέ is thought to reflect the greatest degree of editing because it signals ‘(slight) textual discontinuity’ (Allan [Reference Allan and Lallot2011a: 41], referring to Bakker [Reference Bakker1993]). Allan (Reference Allan and Lallot2011a) presents some data from Thucydides suggesting the present has a slight predilection for the particle καί over δέ, which he relates to the principle of immediacy.Footnote ⁸³ I will focus on these two particles here, hypothesising that the use of καί as a connective increases the odds of the main clause verb being marked with the present. The results are listed in Table 2.6.

Table 2.6Tense and δέ versus καί

^a B = .408 (constant = -.896), standard error = .210.

An effect is observed (odds ratio 1.503), but this is still relatively uncertain (p = 0.052).

Third, I hypothesise that the absence of a connective particle increases the odds of the present for preterite being used in comparison with and-connection. The results are listed in Table 2.7.

Table 2.7Tense and ‘and-connection’ versus asyndeton

The observed counts are identical to the expected counts (therefore, there is no need to run a regression analysis). However, there is more to be said. In almost all cases where no connective particle is used, the effect is mitigated by certain factors. There is often a temporal connective (εἶτα [‘next’]) or an anaphoric pronoun (ἐνταῦθα [‘at that point’]). In other cases, discourse considerations play a role: asyndeton is natural at the very beginning of a narrativeFootnote ⁸⁴ or in a further explication of something that has already been narrated.Footnote ⁸⁵

However, there is a marked usage of asyndeton that is observed with the present for preterite. One instance is discussed by Allan (Reference Allan, Bakker and Wakker2009: 179):

Allan argues that ‘the speed of the actions is iconically expressed by the asyndetic juxtaposition of σκάπτει [‘digs up’] μοχλεύει [wrenches open]’. A similar usage is found in the following example:

I have found no instances of the aorist in such ‘asyndetic juxtapositions’, as Allan calls them. While this finding does not tell us much from a quantitative perspective, it is at least suggestive of a relationship between present tense usage and this particular form of iconic discourse connection.Footnote ⁸⁶

Finally, I turn to prospective discourse connection. The particle μέν (generally untranslatable in English) signals that the host assertion should be viewed in connection with a following one, which is then introduced by δέ.Footnote ⁸⁷ I hypothesise that the presence of the particle μέν decreases the odds of the present for preterite being used. The results are listed in Table 2.8.

Table 2.8Tense and the particle μέν

^a B = -.795 (constant = -.814), standard error = .555.

The distribution suggests a substantial effect (odds ratio 0.451), but there are very few cases of μέν in total, so that this is statistically uncertain (p = 0.152).

In conclusion, the general hypothesis explored in this section was that simplicity in discourse connection increases the odds of the present being used instead of the preterite. While none of the individual results amounted to much in terms of statistical significance, the overall impression is important as well: the results were either positive (Tables 2.6 and 2.8) or neutral with respect to the hypothesis (Tables 2.5 and 2.7 – ignoring the slight effect in Table 2.5, which is basically a 50/50 proposition). Moreover, incidental observations suggest that a particularly vivid type of asyndeton may be associated with the present for preterite.

2.4.3 Sentence Complexity

According to the general hypothesis explored in this section, sentence complexity is expected to decrease the odds of the present being used in the main clause instead of the preterite. I make a distinction between participial clauses and subordinate clauses. In the group of participial clauses, I have included both conjunct participial clauses (where the subject is coreferential with the subject in the main clause) and absolute participial clauses (where the subject is different from the subject in the main clause). I do not present all the counts (the categories range from 0 to 3),Footnote ⁸⁸ but simply note that the results are negligible (odds ratio 1.099, p-value 0.481).Footnote ⁸⁹

Next, I considered temporal subordinate clauses introduced by ὅτε (‘when’), ἐπεί (‘when’), ὡς (‘as’). As the number of such subordinate clauses rarely exceeds 1, I have created only two categories: simple (no temporal subordinate clause) or complex (1 or more temporal subordinate clauses). The results are listed in Table 2.9.

Table 2.9Tense and temporal subordinate clauses

^a B = -.479 (constant = -.794), standard error = .316.

Some effect is observed (odds ratio 0.62) but it is statistically uncertain (p = 0.13).

2.4.4 Position of the Verb

Kroon (Reference Kroon, Sawicki and Shalev2002: 195) lists ‘frequent initial position of the verb’ as one of the markers of the immediate narrative mode in Latin. In my view, this makes sense for Classical Greek as well. When the information conveyed by the verb is focal (i.e., central to the assertion), it usually precedes other focal constituents. This means that, when a sentence is made up of entirely new information, the verb will be in first position (Matić [Reference Matić2003: 582–8]). Topical (presupposed) information may either precede the verb or follow it. I argue that the placement of the verb before topical information reflects the urgency of a crisis situation.

Let me illustrate these principles with two contrastive examples. In Aristophanes, Birds 496–7, we read κἄρτι προκύπτω | ἔξω τείχους καὶ λωποδύτης παίει ῥοπάλῳ με τὸ νῶτον (‘and I just stoop forward out of the city walls and a mugger hits me in the back with a club’). The mugger mentioned in the second clause is a new referent in the discourse, but he is here construed as topic: λωποδύτης (‘mugger’) stands at the initial position of the clause, before the verb παίει (‘hits’), which carries focal information.Footnote ⁹⁰ Information structure here is not iconic of actual experience. In reality, the speaker must first have felt the blow, and only later realised who dealt it; but as we process the discourse, we identify the perpetrator before we learn what he did. Contrast the following example:

Here, the verb παίει (‘hits’) occupies the very first position in the main clause, which reflects the character of the original experience: first the blow is felt, then the other aspects of the situation are observed or deduced.Footnote ⁹¹

My hypothesis, then, is that placement of the verb in the initial position of the main clause increases the odds of the present for preterite being used. The definition of the ‘main clause’, however, is problematic: the intervention of subordinate clauses renders constituents ‘extra-clausal’ (Matić [Reference Matić2003: 580–2]; Allan [Reference Allan2014]).Footnote ⁹² I have simplified the matter by including complex sentences only when the main clause assertion precedes all participial and subordinate clauses ([23] was thus excluded). The cases that were included were coded for whether the position of the verb was initial or non-initial.Footnote ⁹³ The results are listed in Table 2.10.

Table 2.10Tense and the position of the verb

^a B = 1.016 (constant = -1.165), standard error = .233.

A substantial effect is observed (odds ratio 2.762), and the intrinsic probability of the correlation being due to chance is quite low (p < 0.001).

2.4.5 Voice

Coding for voice in Classical Greek is complicated by the fact that semantic categories are not neatly mapped onto morphological categories.Footnote ⁹⁴ In the present tense, there is a morphological distinction between an active and a middle voice. ‘Passive’ is only one of several semantic categories covered by the polysemous middle voice. For the aorist, there is a separate morphological category we call ‘passive’. However, this category encompasses not only true passives (ὤφθην [‘was seen’]) but also intransitive verbs (ἥσθην [‘feel joy’]) and intransitive counterparts of causative verbs (active ὥρμησα [‘set in motion’] with ‘passive’ ὡρμήθην [‘rushed’]).

Establishing a baseline category of ‘active’ forms is problematic for similar reasons. Not all morphologically active forms have corresponding passives: some are intransitive (θνῄσκω [‘die’]), some are causatives, with the passive having intransitive meaning. The matter is further complicated by the existence of an ‘internal accusative’ in Classical Greek that can be used with intransitive verbs. An example would be ναυμαχεῖν ναυμαχίαν, literally ‘sea-fight a sea-fight’. Such constructions are sometimes passivised.

In light of such difficulties, I do not present a statistical analysis but limit myself to some observations. In the ‘present’ category, true passives are rare. There are only two cases where an active construction would have been possible. First, Sophocles, Antigone 422–3 καὶ τοῦδ’ ἀπαλλαγέντος ἐν χρόνῳ μακρῷ, | ἡ παῖς ὁρᾶται (‘and when this passed after a long time, the child is seen’). Here, ‘we see the child’ would have been completely natural. Second, Euripides, Alcestis 183–4 πᾶν δὲ δέμνιον | ὀφθαλμοτέγκτῳ δεύεται πλημμυρίδι (‘the whole bed is wetted by a flow of tears’). Alcestis falls on her bed, crying, and thereby makes it wet.Footnote ⁹⁵ Other cases are more difficult. For example, at Sophocles, Electra 746–7 we read σὺν δ’ ἑλίσσεται | τμητοῖς ἱμᾶσι (‘and he is rolled up in the reins cut from leather’). The event designated here occurs spontaneously (Orestes falls out of his chariot and is caught in the reins) and an active construal would be extraordinary.Footnote ⁹⁶

While passive present for preterite forms are rare, this does not mean much because the same is true for the aorist. A particularly marked instance is Sophocles, Ajax 729–30 ὥστ’ ἐς τοσοῦτον ἦλθον ὥστε καὶ χεροῖν | κολεῶν ἐρυστὰ διεπεραιώθη ξίφη (‘so that they came to a point were even swords were drawn from their sheaths with hands’). An active construal (‘they drew’) would have been much more natural here.Footnote ⁹⁷ But again, other cases are not as straightforward. Consider Euripides, Hippolytus 1246–7 ἵπποι δ’ ἔκρυφθεν καὶ τὸ δύστηνον τέρας | ταύρου λεπαίας οὐ κάτοιδ’ ὅποι χθονός (‘the horses and the wretched bull-beast were hidden somewhere in the rocky ground’). The agent here is unknown, and we may well understand ‘were hidden’ as ‘disappeared’.Footnote ⁹⁸

In conclusion, there is no clear evidence that a passive construal influences tense usage. There are simply too few unambiguous instances of truly passive verb forms overall.

2.4.6 Number

I hypothesise that a plural subject has a negative effect on the odds of the present for preterite being used. The results are listed in Table 2.11.Footnote ⁹⁹

Table 2.11Tense and number

^a B = -1.020 (constant = -.696), standard error .288.

A substantial effect can be observed (odds ratio 0.361), and the intrinsic probability of the correlation being due to chance is quite low (p < 0.001). This supports the idea of the mimetic use of the present for preterite being associated with grammatical simplicity.

2.4.7 Verbal Simplicity

We have been concerned with verb use earlier (Section 2.2.1), but here the issue is not what kind of event is designated but how it is construed by the verb phrase. The hypothesis is that the use of uncommon expressions to designate simple events decreases the odds of the present for preterite being used. A proper quantitative analysis would require a complex coding procedure, so I will only limit myself to discussing some suggestive examples here.

To begin with, consider the following passage:

There is a contrast here between the present παίω (‘hit’) and the aorist καθίκετο (‘came down on’). A search in the TLG shows that the word καθικνέομαι (‘come down on’) is uncommon in the literature from the Classical period, and example (24) is the only instance where it is used of a physical strike. The verb παίω (‘hit’) is the most direct expression for the conceptual content of someone hitting another, as in Demosthenes, Against Evergus and Mnesibulus (47) 38 εἰσιόντος δέ μου παίει πὺξ ὁ Θεόφημος τὸ στόμα (‘as I entered, Theophemus hits me in the mouth with his fist’). (Note also the sentence-final position of the aorist verb in example [24]; compare Section 2.4.4.)

Here is another example (compare Section 2.2.2):

The event designated by the present ἀναφέρει (‘brings)’ is highly concrete and described in simple terms. Eteocles’ action with the spear, by contrast, is described in a figured manner. The verb καθῆκεν (‘sent down’) does not reflect the violent nature of the act. Similarly, to say that Eteocles ‘fitted in’ (ἐνήρμοσεν) his spear in Polynices’ spine is a rather complicated way of expressing what happened. The straightforward alternative would be ὠθέω (‘drive through’), which we actually find in the present for preterite later in the same play:Footnote ¹⁰⁰

Such examples are, of course, merely suggestive, but they do indicate that verbal simplicity may be a factor that allows us to account for instances of the preterite that are otherwise difficult to explain given the ‘mimetic potential’ of the designated events.Footnote ¹⁰¹

2.4.8 Conclusion

I have argued in this section that the linguistic construal of the conceptual content conveyed by the narrator is mimetic in so far as the discourse structure, syntax and word choice facilitate a ‘natural’ kind of processing that mirrors the processing of immediate experience. I then explored the hypothesis that this type of narrative mimesis has a positive influence on the odds of the present being used instead of the preterite. I discussed seven relevant parameters and presented data from the corpus pertaining to each. With respect to discourse time versus story time (Section 2.4.1) and verbal simplicity (Section 2.4.7), I was only able to discuss some suggestive examples. In the case of voice (Section 2.4.5), quantitative analysis was problematic as well, as it turned out that true passives were rare. As far as quantitative analysis was possible, substantial effects were observed for verb position (Section 2.4.4) and number (Section 2.4.6). The data for discourse connection (Section 2.4.2) and sentence complexity (Section 2.4.3) yielded only hints of an effect here and there.

I think the main reason for the absence of strong evidence in certain cases is that narrative in drama is generally mimetic in character. For example, temporal subordinate clauses, passives and particles marking complex discourse relations are rare in the corpus. This makes differences between the tenses with respect to these parameters slight. Nevertheless, I conclude that, overall, the results lend some (if not overwhelming) support to the hypothesis that mimesis in linguistic construal positively correlates with present for preterite usage.

2.5.1 Communicative Dynamism

The preceding discussions may have given rise to the impression that I view the mimetic use of the present for preterite as the necessary outcome of some formula where all relevant variables are measures of narrative mimesis – as if a certain level of narrative mimesis were both necessary and sufficient conditions for the use of the present for preterite. There are, however, two sides to the equation. Narrative mimesis certainly facilitates the construal of the narrated events as being part of the present, but on the other hand, the use of the present tense is itself part of a strategy to impose such a construal on the communicative situation. This means that the rhetorical concerns of the narrator also need to be taken into consideration when accounting for tense-switching in discourse. My argument in this section is that the influence of narrative iconicity on tense-switching is moderated by the degree of communicative dynamism in the narrated events (McNeill [Reference McNeill1992], [2005]).

Communicative dynamism concerns the newsworthiness or relevance of the designated events in the context of the communicative situation. I distinguish between three parameters. First is ordinariness versus unexpectedness. Events that evolve according to a certain protocol, custom or routine are low in communicative dynamism. Transgressive acts, unexpected developments and miraculous phenomena are the opposite. Second, events are high in communicative dynamism when they are directly relevant to the present communicative situation.Footnote ¹⁰³ The third parameter is whether the designated events involve an increase or decrease in narrative tension. In the overall structure of narrative, the narrative tension culminates in the peak and is released in the following resolution.Footnote ¹⁰⁴

The importance of communicative dynamism for tense-switching can be related to narrative mimesis through the phenomenology of attention, as I explained in the introduction to this chapter. In actual experience, our attention may be more or less focused depending on what is happening around us. When something unusual or unexpected happens, our attention is heightened – especially if the event poses some kind of threat. If narrative is mimetic of actual experience, then our engagement with the narrative will similarly be heightened when the designated events are high in communicative dynamism.

I will discuss three contrastive examples to illustrate these points here; more detailed and extensive analyses of text passages will be presented in Section 2.6. The three cases presented here are centred around particular verbs that are marked with the present in one context but with the aorist in another. One point I should make is that it is impossible to find contrastive pairs that are perfectly equivalent with respect to narrative mimesis and differ only in the degree of communicative dynamism. In fact, the two often seem to go hand in hand, so that salient information is presented in a more strongly mimetic manner.Footnote ¹⁰⁵

My first contrastive pair involves two passages where characters jump up (the Greek phrase is ὀρθὸς ἀνᾴσσω) after waking up. The context of the first passage is as follows: Hector has summoned King Rhesus and his Thracian army to aid Troy in the war against the Greeks. At night, the Thracian camp is attacked by Odysseus and Diomedes. Rhesus’ charioteer witnessed and survived the attack but is unable to identify the perpetrators. He tells how, when he woke from a nightmarish dream, he found the camp was under attack:

The attack on the camp is an extremely transgressive act, and it comes as a complete surprise to the charioteer. From the moment the charioteer wakes from his portentous dream, the narrative tension rises, until it reaches its climax when he is struck down by one of the attackers. From a larger narrative-structural perspective, this passage constitutes the main crisis (peak) of the narrative. The main point of the narrative is that the camp has been attacked and that the Thracians, including Rhesus, have been slain. The preceding passages (762–78: the charioteer feeds the horses; 779–86: he has a troubled dream) prepare for the main conflict narrated in the passage cited above. After this, the imperfect in ἵεσαν φυγῇ πόδα (‘sent their foot to flight’, 798) marks typical resolution material: the perpetrators got away unhindered.

In the following contrastive example, a servant of Pentheus, king of Thebes, describes the actions of the women who left Thebes in a frenzy to serve Dionysus:

The phrase ἀνῇξαν ὀρθαί (‘jumped straight up’) in 693 mirrors ὀρθὸς δ’ ἀνᾴσσω (‘and I jump straight up’) in example (27). In the present instance, there is an almost complete absence of narrative tension. Where the charioteer in (27) wakes up in a state of distress due to a bad dream, the bacchants wake up after ‘refreshing sleep’ (692 θαλερὸν ὕπνον). The charioteer in (27) jumps up in agitation, but when the bacchants rise, they are a ‘marvel of orderliness to behold’ (θαῦμ’ ἰδεῖν εὐκοσμίας). In (27), the narrative tension culminates soon after the charioteer has risen from his sleep. Here, by contrast, the bacchants simply put their attire in order.

In (27), high communicative dynamism goes hand in hand with strong narrative mimesis. The events marked with the present are all highly concrete and intense: waking up in a state of agitation (787 ἐξεγείρομαι [‘wake up’]), hearing men dying (789 κλύω [‘hear’]), feeling blood spurting against one’s body (791 βάλλει [‘hits’]), jumping up (792 ἀνᾴσσω [‘jump up’]), getting hit (794 παίει [‘hits’]), falling (797 πίπτω [‘fall’]). These are traumatic experiences that will stand out vividly in the speaker’s memory. Moreover, discourse time is highly iconic to story time. In (28), the designated actions are concrete as well, but less forceful or intense (695 καθεῖσαν [‘sent down’], 696 ἀνεστείλαντο [‘tucked up’], 698 κατεζώσαντο [‘girded’]). Also, these actions take more time to occur than the punctual events marked with the present in (27), especially as they are carried out by different individuals (for the effect of a plural subject on tense usage, see Section 2.4.6).

My second contrastive pair concerns two instances of the verb ἐκραίνω (‘scatter out of’), referring to the shocking act of smashing someone’s skull against a rock. In Sophocles’ Women of Trachis, Hyllus tells the story of how his father Heracles was overcome by frenzied pain during a sacrifice. In his madness, Heracles throws his servant Lichas against a rock:

The present forms mark a highly shocking act and an unexpected turn of events in the story. Compare the following passage from Euripides’ Cyclops, where Odysseus tells how the cyclops feasted on his companions:

In the previous passage (382–95), we have been told of the elaborate preparations made by the cyclops before he starts cooking. The preterite is consistently used to mark these actions. Now that everything is ready, the cyclops proceeds according to plan. While the event designated by the aorist ἐξέρρανε (‘scattered out’, 402) marks an act that is, in itself, shocking, in the present context it is merely one step in the preparation of the cyclops’s meal. There is no real narrative tension because Odysseus and his companions are completely helpless. Salient developments start only at 411–12, where Odysseus figures out a plan to escape (note the present προσφέρω [‘bring’]) in 412).Footnote ¹⁰⁶

With respect to narrative mimesis, the description in (30) is rather dry and ‘objective’ compared to that in (29). The phrase λευκὸν μυελόν (‘white marrow’) to designate the object in (29) is more vivid than the designation ἐγκέφαλον (‘brain’) in (30). In (29), the word κόμης (‘from his hair’) adds another vivid detail, and the narrator elaborates with the gruesome phrase μέσου | κρατὸς διασπαρέντος αἵματός θ’ ὁμοῦ (‘as the head was shattered and blood was mixed in’). Also, the juxtaposition of the main clause verbs ῥίπτει (‘throws’) and ἐκραίνει (‘scatters out of’) is more iconic of actual experience than the hypotactic construction παίων … ἐξέρρανε (‘smashing … he scattered out of’) in (30).Footnote ¹⁰⁷

Finally, let us consider two instances where the expression πέπλους ῥήγνυμι (‘tear one’s robes’) is used. In Aeschylus’ Persians, Atossa, the mother of king Xerxes, tells of a distressing dream she had concerning her son. In her dream, two women, one Persian, one Greek, were subdued by Xerxes and yoked to a chariot. One of the women was obedient, but the other broke free:

This cluster of six main clause present for preterite forms (note also the present ὁρᾷ (‘sees’) in the subordinate clause in 198) is exceptional.Footnote ¹⁰⁸ The episode is full of violent action and is highly distressing to Atossa, as it is a sign of her son’s fate in Greece. In this context, Xerxes’ gesture of tearing his robes is poignant: he has suffered defeat at the hands of the Greeks and is a failure in the eyes of his father. This foreshadows the parts in the play where the ghost of Darius offers his perspective on Xerxes’ actions (681–842) and where Xerxes appears, lamenting his defeat (from 908 to the end).

Contrast the following passage, where Polyxena gets ready to be sacrificed by the Greeks to placate the soul of Achilles:

Unlike Xerxes, Polyxena has become resigned to her fate: she announces that she is prepared to die voluntarily in 546–52. There is a consistent emphasis on Polyxena’s remarkable composure throughout the ritual: even as she died, she ‘took great care to fall with decorum’ (569 πολλὴν πρόνοιαν εἶχεν εὐσχήμων πεσεῖν). In this context, then, the event of Polyxena tearing her clothes does not carry particular weight; it is merely part of the necessary preparations for the sacrifice.

2.5.2 Conclusion

In this section, I have argued that communicative dynamism acts as a moderating influence on the effect of narrative mimesis on tense-switching. While narrative mimesis facilitates the construal of past events as presently accessible, the newsworthiness or relevance of the designated events influences the propensity of the narrator to highlight this construal with the present tense. However, as the contrastive pairs discussed here illustrate, the two parameters are difficult to tease apart: a mimetic style of narration seems to be preferred precisely in those parts that are especially noteworthy and important to the discourse.

2.6.1 Sophocles, Electra 893–908 and Euripides, Electra 509–19

Sophocles’ and Euripides’ Electra plays both revolve around the plot of Orestes’ return to the house of Agamemnon. After Agamemnon returned from Troy, he was murdered by his wife Clytaemnestra, whose new husband Aegisthus usurped the throne. Agamemnon’s son Orestes was banished, while his daughter Electra remained behind. In Sophocles’ version, Electra still lives in the palace, but in Euripides’ version, she has been married out to a farmer and lives in the country.

In both plays, someone visits the tomb of Agamemnon and sees that sacrificial rites have been performed for the dead king. This person tells Electra what he/she saw. In Sophocles’ version, the narrator is Chrysothemis, Electra’s sister; in Euripides, it is an old man, who was once Agamemnon’s nurse (Chrysothemis has no part in this play). The two narratives make for a striking contrast with respect to tense usage, as Chrysothemis predominantly uses the present tense, while the nurse sticks to the preterite.

Let us begin with Sophocles’ version. Chrysothemis returns from the tomb of her father Agamemnon in a jubilant state, for she is positive that Orestes has returned (871–91). Electra is skeptical, but listens to Chrysothemis’ tale:

Chrysothemis’ narrative iconically conveys the character of her actual experiences. Finglass (Reference Finglass2007: 380) calls the narrative ‘engaging’ and makes a number of observations that square well with my mimetic account of the present for preterite.

First, ‘the stress on [Chrysothemis’] personal autopsy of the events’. We find many verbs of ‘seeing’ or ‘looking’ throughout: 894 ὁρῶ (‘see’), 897 ἰδοῦσα (‘seeing’), 897 περισκοπῶ (‘look around’), 899 ἐδερκόμην (‘saw’), 900 ὁρῶ (‘see’), 902 εἶδον (‘saw’). There is also simulated perception, as the mental image of Orestes strikes Chrysothemis (903 ὄμμα [‘image’] 904 ὁρᾶν [‘beholding’]). This consistent emphasis on perception heightens the mimetic quality of the narrative.

Second, ‘the emphasis on the discovery as a process’. As we follow the narrative, ‘we observe Chrysothemis’ increasing emotion, as she moves from surprise (897) to suspicion (897–8), then to joy (902–6)’ (Finglass [Reference Finglass2007: 380]; compare de Jong [1991: 35–8] on similar techniques in Euripidean messenger speeches). Again, this strengthens our empathy with the narrator, inviting us to simulate her emotions so that we can recreate the narrated experiences for ourselves.

Third, Finglass (2007: 380) argues that ‘[s]ome elements in her speech have a colloquial edge to them’. He points to the use of the present for preterite, but also to an aspect of sentence construction. At 897, the participle ἰδοῦσα (‘seeing’) picks up the main verb ὁρῶ (‘see’) in 894; similarly, at 902 the subordinate clause ὡς εἶδον (‘at the moment that I saw it’) picks up ὁρῶ (‘see’) in 900. This type of construction is considered a feature of ‘unsophisticated’ or ‘unelevated’ narrative. Finglass argues that this colloquial character ‘conveys the infectious excitement of the narrator’. This can be related to linguistic simplicity as narrative mimesis (Section 2.4).

Now for the individual present forms. The use of the present to mark sights (ὁρῶ [‘see’] in 894 and 900) is typical, as simulated perception is highly iconic of actual perception. In 902, the verb ἐμπαίει (‘strikes’) refers to a mental image, which is already simulated perception in itself; so the simulation potential may be even higher here, as a simulation of simulated perception will be even more iconic than a simulation of actual perception. Note, furthermore, that there is a strong affective component: Chrysothemis construes the mental vision as something which ‘strikes’ her, and at this point in the narrative she highlights her misery with the word τάλαινα ‘wretched’ (rendered ‘ah!’ by Lloyd-Jones).

The emotional impact of the realisation that Orestes has returned is further emphasised in lines 905–6. The present forms in the balancing statements (δυσφημῶ [‘utter ill-omened words’] and πίμπλημι [‘fill’]) must be taken together as one whole, with the weight lying on the positive statement. That is, the negated form is in the present mostly because it anticipates the following positive statement.Footnote ¹¹⁰ The feeling of being overcome with emotion is concrete and intense.Footnote ¹¹¹

The remaining present form is περισκοπῶ (‘look around’) in 897. This is a somewhat unusual instance, as the verb designates an activity (as Boter [Reference Boter2012] rightly points out; see Introduction, Section I.4.5). The implicit aspectual construal here is most probably imperfective, as suggested by the imperfect ἐδερκόμην (‘saw’) in the following narrative subordinate clause. My explanation for the use of the present here is the strong simulation potential of the verb. Looking around is a concrete bodily activity (moving the head and the upper body in different directions), and we may well imagine that at this point the narrator would mimic the act of looking around.

There are two preterites in this narrative. First there is the aorist ἔσχον (θαῦμα), lit. ‘caught (surprise)’. This is an ingressive aorist of a stative verb; the present is generally excluded here.Footnote ¹¹² The imperfect προσεῖρπον (‘crept’) is chosen, I think, because Chrysothemis wishes to highlight the imperfective aspectual construal: this slows down the narrative and creates suspense, as we wonder what she will find next. The effect is reinforced by the adverb ἆσσον (‘closer’), which does not signal a definite endpoint.

Now let us compare Euripides’ version, where it is Agamemnon’s old nurse who tells the story:

The aorist ἦλθον (‘went’) is typical at the very beginning of the story when there is no subordinate clause to set the scene. In Chrysothemis’ narrative, we find this same verb in such a subordinate clause (893). The actions of the nurse at the grave (ἔκλαυσα [‘wept’], ἔσπεισα [‘poured’], ἀμφέθηκα [‘put around’]) are customary: there is nothing remarkable about a servant paying respects to his former master at his grave. Also, the verb phrases κλαίω (‘weep’) and σπονδὰς σπένδω (‘pour libations’) designate activities; and putting myrtle branches around the tomb (ἀμφέθηκα [‘put around’]) is a durative action. This makes the narrative tempo less iconic here than in Chrysothemis’ narrative.

The most interesting case is that of ἐσεῖδον (‘saw’). In contrast to Chrysothemis, the old man does not mark this remarkable discovery with a present tense.Footnote ¹¹³ Where Chrysothemis is sure she has seen signs of Orestes’ return (907–8), the old man only tentatively suggests this possibility (ἴσως [‘perhaps’] in 518).Footnote ¹¹⁴ In fact, lines 518–19 and the following discussion of the possible evidence for Orestes’ return (520–44) have been regarded as suspect by some scholars (see Roisman and Luschnig [Reference Roisman and Luschnig2011] ad 520–44). If the passage turned out to be a forgery, this would only underscore my point that the scene at the grave of Agamemnon has less significance for the old man than for Chrysothemis.

The difference in communicative dynamism of the discovery of the tokens of Orestes in the two contexts is reflected in the use of syntax by the respective narrators. The old man places the verb at the very end of the clause:

There follows an elaboration of two more objects (the blood and the hair), but this makes little difference. The point is that the verb ἐσεῖδον (‘saw’) is placed after both an object (οἶν μελάγχιμον πόκῳ [‘a black lamb’]) and an object complement (σφάγιον [‘slaughtered’]). The present seems generally dispreferred with this kind of word order.Footnote ¹¹⁵ In my view, low information status for the verb reflects narratorial mediation and a measure of distance (compare Section 2.4.4). Chrysothemis, by contrast, places the verb ὁρῶ (‘see’) at the very first position of the main clause in 894 and almost in first position in 900. The early placement of the verb conveys an excited tone.

The final main clause verb in the narrative, ἐθαύμασα (‘wondered’), designates a state (‘ingressive aorist’; see Appendix, Section A.2.1).

To conclude, the difference in tense usage between Sophocles’ and Euripides’ versions of the discovery of the tokens of Orestes corresponds to the differences in character between these narratives. Chrysothemis, elated at the certain return of Orestes, construes a narrative that conveys the feeling of actual experience through its ‘stress on personal autopsy’ and the ‘emphasis on the discovery as a process’ (Finglass Reference Finglass2007: 380). The use of the present for preterite as the main narrative tense highlights the simulative character of the narrative. The old man in Euripides’ version, by contrast, tells his story in a state of dejection. The first half of his narrative is merely concerned with the ritual formalities he performed at the grave of his former master (509–12). The discovery of tokens of a sacrifice at the site cause him to wonder (513–17), but he is cautious in assessing their significance. The consistent use of the preterite fits the inhibited character of the narrative.

2.6.2 Euripides, Heracles 963–1012: Heracles’ Madness

The messenger narrative in Euripides’ Heracles (922–1012) describes how Heracles kills his sons in a fit of madness. Heracles has just killed Lycus, the usurper of the throne of Thebes, and is about to perform a purification sacrifice. Suddenly, he is overcome with madness. The idea occurs to him that he should first kill Eurystheus, the man who sent him on his labors, and only then purify himself. He then imagines himself riding a chariot to Mycenae; in fact, he is only moving back and forth through the house.

I start my discussion at the point where Heracles imagines having arrived at the city of Eurystheus, bent on killing the man and his children (who are, of course, nowhere near):

Heracles’ father tries to reason with him. The use of the present to introduce an anxious plea (ἐννέπει [‘says’]) is typical; compare βοᾷ (‘shouts’) in 975 and αὐδᾷ (‘says’) in 988 (see below, where I adduce another parallel).Footnote ¹¹⁶ Heracles pushes his father away. The present ὠθεῖ (‘pushes’) denotes concrete and forceful manual action.Footnote ¹¹⁷ Heracles then makes ready his bow and quiver to use against his children. The present in εὐτρεπῆ σκευάζεται (‘makes ready’) is not strongly mimetic. Making one’s equipment ready does involve manual action, but no force is involved, and the action is durative. However, in the present context, the designated event is high in communicative dynamism, as it means that Heracles now turns against his own children. At this point the situation turns into a full-blown crisis.

Heracles’ children seek refuge. The aorist ἔπτηξε (‘cowered’) merely expands on information already given by the imperfect ὤρουον (‘rushed’) at 972. I think the lack of specific identification of the children is also a factor here; they are simply labelled 972 ὁ (‘one’), 973 ὁ (‘another’), 974 ἄλλος (‘another’). The aorist is preferred in such contexts.Footnote ¹¹⁸

Heracles’ mother ‘cries out’ (βοᾷ), trying to get Heracles to come to his senses. The cry is seconded by the father and the servants (another present βοᾷ [‘cry out’]). Again, the present highlights an anxious plea, and the verb is strongly mimetic (Section 2.2.1). Note also the initial position of the verb in both cases (Section 2.4.4).

In the next part of the narrative, Heracles kills one of his children:

Heracles shoots his child in the liver; the present βάλλει (‘shoots’) marks this violent and shocking act. Dying, the child ‘drenched’ (ἔδευσεν) the pillars (with blood). It is instructive to compare the aorist here with two instances of the present referring to similar events. In the first, Agamemnon dies at the hands of his wife Clytaemnestra:

In the second, Haemon commits suicide in the arms of Antigone:Footnote ¹¹⁹

In both these cases, the spurting out of blood is construed as a violent physical action: the dying person is said to ‘eject’ (ἐκβάλλει) a stream of blood or to ‘hit’ (βάλλει) another with his blood. Moreover, the description as a whole is quite graphic in both instances. At Heracles 980, by contrast, the gruesome details are only implied. Heracles’ child is said to ‘drench’ the pillars, but there is no mention of actual blood, let alone a graphic description such as ἐρεμνῇ ψακάδι φοινίας δρόσου (‘a black shower of gory dew’) in (33) or ὀξεῖαν ῥοὴν φοινίου σταλάγματος (‘a sharp stream of oozing blood’) in (34). This inhibition on the part of the narrator is reflected in the use of the aorist ‘drenched’ (ἔδευσεν) to mark the event, rather than the present.Footnote ¹²⁰

Having killed his child, Heracles celebrates his victory: this has no impact on the narrative dynamic. Moreover, ἠλάλαξε (‘shouted in triumph’) is an activity verb (Section 2.2.1). The form ἐκόμπασεν (‘uttered a boast’) introduces direct speech, but Heracles’ words lack the agitation or consternation of the speakers whose words are marked with the present tense in this narrative (or elsewhere; see Section 2.3.1).

Next, Heracles approaches his second victim:

The phrase φθάνει … προσπεσών (‘throws himself at [Heracles] before [he could shoot]’) requires some explanation. The auxiliary verb φθάνω signals that the action designated by the participle was carried out before someone else could prevent it. Thus, it inherently carries with it connotations of speed and decisiveness, and consequently it has a strong predilection for the present for preterite in general (compare the present with the same verb at 996, see below).Footnote ¹²¹ In this particular instance, the action that is actually carried out is designated by the verb προσπίπτω (‘throw oneself at’), designating forceful bodily movement.Footnote ¹²²

Next, the present αὐδᾷ (‘says’) marks the son’s supplication for his life. A striking parallel is Pentheus’ appeal to his frenzied mother Agave in Euripides’ Bacchants:

Both Pentheus and Heracles’ son beg for their life, both address someone who is in a state of madness and does not recognise them and both touch their aggressor as a token of supplication. Also, the present αὐδᾷ (‘says’) in Heracles 988 is in inquit-position, so that the influence of the mimesis of direct speech representation on the tense of the reporting verb may be stronger (compare Section 2.3.2).

The actual killing of the second child is marked with two aorists: καθῆκε (‘sent down’) and ἔρρηξε (‘smashed’). This seems unexpected, as the killings are violent actions and constitute the essential part of the story. Allan (Reference Allan, Bakker and Wakker2009: 195) comments on this and suggests that the second and third killings are not of special interest to the addressees (‘less tellable’) because they may be expected after the first killing. I think this is a relevant point, but I think it needs to be supplemented with additional linguistic arguments.

I argue that verbal simplicity is the central factor here (Section 2.4.7). The phrase ‘send (something) down (on someone)’, is not a straightforward reference to an act of hitting; that would be παίω (‘hit’) or κόπτω (‘hit’). In the first place, the complement structure of καθίημι (‘send down’), consisting of an object and a prepositional phrase, is more complex than that of παίω (‘hit’) or κόπτω (‘hit’). Second, the verb καθίημι (‘send down’) lacks the inherent connotation of forceful movement that is present in the alternatives just mentioned. The use of a circumlocutory and less forceful expression signals a measure of narratorial distance here that is at odds with the principle of mimetic narration. This may well tie in with Allan’s observation that the killing itself is no longer as high in communicative dynamism (‘tellable’) as it was in the first instance.

As for ἔρρηξε (‘smashed’), this is not a separate action but rather the consequence of what was narrated in the previous clause. The two verbs καθῆκε (‘sent down’) and ἔρρηξε (‘smashed’) constitute a single sequence, so that continuity of tense usage is natural here.Footnote ¹²³

Heracles now turns against his third child:

The present form χωρεῖ (‘goes off’) marks a new development.Footnote ¹²⁴ This use is typical with verbs of movement in dramatic narrative and may be understood in terms of communicative dynamism (compare 1001 ἱππεύει [‘gallops’]; see Allan [Reference Allan, Bakker and Wakker2009]). This is especially the case when, as here, a strong feeling of danger is involved.

The narrative dynamic intensifies as Heracles’ wife ‘takes (the child) away before (Heracles can kill him)’ (φθάνει … ὑπεκλαβοῦσα) and tries to prevent Heracles from reaching the child by shutting the gates to the inner quarters (κλῄει [‘bars’]). The auxiliary verb φθάνω, again, implies speed and a change in the narrative dynamic. By closing the door, Heracles’ wife presents the protagonist with a physical barrier, which heightens the narrative tension.Footnote ¹²⁵ Heracles breaks down the barrier. The present tense forms σκάπτει μοχλεύει (‘digs up, wrenches open’) denote violent manual action, and, as Allan (Reference Allan, Bakker and Wakker2009: 179) notes, the juxtaposition of the two verbs without a connective particle iconically suggests speed (see Section 2.4.2).

The third killing is marked with an aorist: κατέστρωσεν (‘laid low’). Again, I would like to supplement Allan’s (Reference Allan, Bakker and Wakker2009: 195) explanation in terms of ‘tellability’ with some more detailed observations. My first point pertains to verbal semantics. The verb καταστορέννυμι properly denotes ‘spreading something upon another thing’. The meaning ‘lay low’, of opponents in battle, is found in only four other instances in Classical Greek.Footnote ¹²⁶ As with καθῆκε (‘sent down’) in 993, I argue that the verb used here lacks the inherent connotation of force that is present in its more direct alternative, which is the expression καταβάλλω (‘throw down’).Footnote ¹²⁷ An illustrative case of the present for preterite with that verb is the following:

A second point concerns syntax. The verb κατέστρωσεν (‘laid low’) occupies a late position in the sentence: it is preceded by two object complements, as well as the part of the instrumental dative noun phrase which carries focus (‘with one projectile’). I suggest that the low information status of the verb here reflects the predictability of the killing itself (compare Section 2.4.4), which was the point made by Allan (Reference Allan, Bakker and Wakker2009: 195).

Finally, Heracles turns against his father:

Similarly to χωρεῖ (‘goes off’) at 995, the present ἱππεύει (‘gallops’) marks the beginning of Heracles’ next venture.Footnote ¹²⁸ Moreover, the description of Heracles’ movement as ‘galloping’ more strongly appeals to the motor faculties of the audience than the blander χωρεῖ (‘goes off’).Footnote ¹²⁹

A phantom appears and stops Heracles’ madness. The use of aorists here is surprising from a narrative-structural point of view, as this constitutes a turning point in the story. Perhaps the displacement of the narrator’s focus from Heracles to another entity, together with the element of narratorial mediation conveyed by ἀλλά (‘but’), somehow conspire to lower communicative dynamism here – the present is used again in 1006, where Heracles is the subject – but I readily admit that, without further evidence for such principles, this is rather ad hoc. I do note that the expression ἦλθεν (‘came’) does not denote concrete movement, nor does it evoke perception (note, however, ὡς ὁρᾶν ἐφαίνετο [‘as it appeared to the eyes’] in the same line). The phantom throws a stone at Heracles. The aorist ἔρριψε (‘hurled’) does designate concrete and violent physical action; could it be relevant that the subject is a phantom manifestation of a god, and therefore maybe the connotation of physical force is diminished?Footnote ¹³⁰

The aorists ἔσχε (‘checked’) and καθῆκε (‘sent’) designate the result of the previous action. Moreover, the aorist ἔσχε (‘checked’) belongs to the stative verb ἔχω ‘have’.

Finally, Heracles falls to the ground: πίτνει (‘falls’). The present is always used with this verb in the selected corpus (Section 2.2.1). Note how the narrator vividly adds that Heracles ‘hit his back against a column’ as he fell (πρὸς κίονα | νῶτον πατάξας). After this, the imperfect ἀνήπτομεν (‘bound’) marks the definitive resolution of narrative tension.

In conclusion, the overall narrative is characterised by violence, speed and a feeling of crisis that is conveyed especially strongly through the pleas of the bystanders. Communicative dynamism is extremely high, as there are few things as shocking as a father killing his own children. These features, in my view, account for the high concentration of present for preterite forms here.

In the next and final case study, we will be looking at an extended narrative passage where the present for preterite is entirely absent. Through this contrast, the dynamics of tense-switching will, I hope, become even more apparent.

2.6.3 Aristophanes, Wealth 696–747: Asclepius Heals the God of Wealth

Two men, Chremylus and his servant Cario, take Plutus, the god of wealth, to a sanctuary of Asclepius to be healed of his blindness. When the party returns, Cario tells Chremylus’ wife what happened in an extended narrative (653–747). I will focus here on the second part of this narrative (696–747), where Cario tells how the god Asclepius appeared and treated his patients. (I cut out some exchanges between Cario and Chremylus’ wife that add nothing to the progression of the narrative.)

First, the god makes his entrance, and Cario commits a blunder:

Cario introduces the next event in the narrative with a narratorial comment: ‘I did something ridiculous’ (γελοῖον … τι ἐποίησα). This is an abstract reference to what he actually did.Footnote ¹³² The actual event is marked with the aorist ἀπέπαρδον (‘farted’). In itself, the designated event is highly concrete (a spontaneous bodily process). However, the event has no repercussions on the narrative dynamic: the god Asclepius pays no heed (704) and goes about his business (709). I think that Cario’s embarrassment (707–8) may also have inhibited the use of the present tense (compare my comments on Kramer’s I popped him: Chapter 1, Section 1.6). It is useful to compare the use of the present for preterite to refer to a spontaneous bodily process in Xenophon, Expedition of Cyrus 3.2.9 τοῦτο δὲ λέγοντος αὐτοῦ πτάρνυταί τις (‘as he was saying this, someone sneezes’). Here the sneezer interrupts a speech, and his sneeze is interpreted by those present as a positive sign from the gods.Footnote ¹³³

Next, we have four aorists that designate the reactions of all characters to Cario’s fart: ὑπηρυθρίασε (‘turned red’; Iaso), ἀπεστράφη (‘turned away’; Panacea), οὐδ’ ἐφρόντισεν (‘didn’t even mind’; Asclepius) and ἐνεκαλυψάμην (‘covered myself’; Carion himself). These reactions, again, do not impact the narrative dynamic in any way. The first and the third of these four verb forms are ingressive aorists of stative verbs (ὑπερυθριάω [‘blush’], φροντίζω [‘mind’]). For the other two, close parallels can be found in Euripides’ Medea, where Jason’s new wife reacts to her husband bringing his children into her home:

Ultimately, Jason’s new wife changes her mind and assents to everything Jason wants, so her initial reaction has no real impact on the narrative dynamic. Moreover, these actions, while concrete, do not involve physical exertion. The aorist ἀπέστρεψε (‘turned away’) nicely mirrors ἀπεστράφη (‘turned away’) at Wealth 702, and προὐκαλύψατο (‘covered’) may be compared to ἐνεκαλυψάμην (‘covered myself’) at Wealth 707.

Now the god gets down to business. In the following passage, he applies medicine to his first patient:

First, a servant brings the god the necessary equipment (παρέθηκε [‘placed beside’]). Placing is a concrete action but it does not require physical exertion. Moreover, there are three grammatical objects here: λίθινον θυείδιον (‘a mortar made of stone’), δοίδυκα (‘a pestle’) and κιβώτιον (‘a box’). So, in fact, the servant commits three acts, and the use of only a single verb here goes against the principle of mimetic grammar.Footnote ¹³⁴ With respect to communicative dynamism, this action only prepares for the actual treatment.

Next, the god ‘started to grind’ [ἐνεχείρησε τρίβειν] a cataplasm). Then he applied it to Neocleides’ eyelids (κατέπλασεν [‘plastered’]). With regard to verbal semantics, ἐγχειρέω (‘start to’) is abstract, while the actual process that is started is an activity (τρίβω [‘grind’]). The verb καταπλάττω (‘plaster’) does designate concrete manual action, but without force or violence, and is also durative. Moreover, applying medicine is a scripted procedure. This is reflected in the use of adverbs. We find πρῶτον (‘first’) at 716, and ἔπειτα/εἶτα (‘next’) at 716, 718 and 720. The use of such adverbs suggests that the events occurred according to a certain controlled procedure, as opposed to in a turbulent fashion.Footnote ¹³⁵

Next, Asclepius treats Plutus:

Asclepius continues the tour of his patients, and the aorists continue to refer to protocollary actions: note again the use of the ‘procedural’ adverbs πρῶτα (‘first’, 728), ἔπειτα (‘next’, 729) and εἶτα (‘next’, 732). First, the god ‘sat down next to’ Plutus (παρεκαθέζετο). The verb denotes concrete bodily movement, but the aspect of intensity is absent. The aorists ἐφήψατο (‘felt’), περιέψησεν (‘daubed around’), κατεπέτασε (‘wrapped’) all refer to non-forceful actions. The second of these verbs is an activity verb, and the third, while telic, is durative.

The god summons two snakes. The aorist ἐπόππυσεν (‘clucked’) belongs to those obviously onomatopoeic verbs (such as ἀλαλάζω [‘cry alalai’], ὀτοτύζω [‘cry ototoi’]) that are never marked with the present, presumably because of their atelic actionality (Section 2.2.1). The appearance of two huge snakes would seem to be a remarkable event (note also the reaction of the woman in 734: ὦ φίλοι θεοί [‘oh dear gods!’]), so we might have expected the present here instead of the aorist (ἐξῃξάτην [‘darted out’]). At the same time, snakes were naturally associated with Asclepian healing rituals.Footnote ¹³⁶ Their appearance here causes no danger or conflict: they peacefully carry out the god’s instructions. Still, I admit that, had the present been used, I would have had no qualms about invoking communicative dynamism to account for it.

What is interesting is that there is no peak-marking present form in this narrative. The explanation lies in the particular way in which the climax, the healing of Plutus, is presented. The actual moment he is healed is not narrated. Instead, we are suddenly confronted with the situation in which Plutus has been healed: ἀνειστήκει (pluperfect) βλέπων (‘he was standing, able to see’).

The last part of the narrative recounts the joyful reaction of Cario and the others present at the healing of Plutus:

All this is resolution material, serving to bring the narrative to a natural close after the main climax (the healing of Plutus). The aorist ἀνεκρότησα (‘clapped’) marks a reaction, in typical fashion.Footnote ¹³⁷ The form ἠφάνισεν (‘vanished’) marks the disappearance of participants from the scene after they have fulfilled their narrative function.Footnote ¹³⁸ Finally, the beginning of the new day (διέλαμψεν [‘shone through’]) marks a natural endpoint to the entire episode (the god operates at night).Footnote ¹³⁹

To conclude, the consistent use of the preterite in this narrative reflects the general smoothness of the narrative dynamic. The main line is constituted by a healing ritual, which has a formal character. The god Asclepius does inflict pain on one of his patients, Neocleides (716–25), but this is non-threateningly comic to the observers (note also γελάσας [‘laughing’] in 723). Carion’s fart and the reactions it provokes are secondary to this main line. After the implied peak (the healing of Plutus), the rest is resolution material. All this makes for a stark contrast with the messenger narrative in Euripides’ Heracles discussed in Section 2.6.2, where the action was full of violence, speed and consternation, and the present for preterite was richly used.