1. Ambiguity
It is hard to find a contemporary scholar actually approving sustained ambiguity in the interpretation of ancient texts, especially in the πίστις Χριστοῦ debate. However, just one, and that quite recently, comes to mind: Robert Jewett, discussing ‘faith of’ or ‘faith in’ Christ at Rom 3.22, argues:
Both the subjective or [sic] objective theories as currently presented have loopholes… It may be that a simple association between ‘faith’ and ‘Jesus’…may have been intended…and neither of the strict construals matches what the original audience would have understood. I wonder whether the ambiguity may have been intentional on Paul's part, so as to encompass the variety of tenement and house churches in Rome that may well have been using the formula of πίστις Χριστοῦ with a variety of connotations.Footnote 1
However, Jewett cites no one in support of hearing two or even more senses together, and on the whole, the guild of biblical scholars would appear to dislike ambiguity, should it appear, or seem to, in the ancient texts, and especially in the πίστις Χριστοῦ debate. It is ‘unfortunate’ (Morna Hooker), ‘dangerous’ (Paul Achtemeier), and needs to be ‘disambiguated’ (Barry Matlock): in each instance the genitive is either objective or subjective, certainly not both.Footnote 2 There is no enquiry into any possible relevance of ancient semantic theory and practice.
To be sure, in narrative (e.g. in the Fourth Gospel) ambiguity may regularly be allowed as intentional and admissible; yet, once two clear senses of a lexeme are recognised, confusion for discerning readers is happily dispelled.Footnote 3 Apart from John, Lauri Thurén has discerned a persuasive ambiguity in 1 Peter's combination of exhortation and encouragement (‘become what you are’).Footnote 4 Mark D. Given has argued (and cogently) for a deliberate employment of ambiguity by the Lukan Paul of Acts 17, and by the authorial Paul of the Corinthian and Roman letters. (In passing, Given glanced at what is the main theme of this paper, pervasive polysemity, rather than deliberate and even precise ‘two-sense’ ambiguity, but did not pursue it.)Footnote 5 We return to Given below.
Todd Still might seem recently to have come fairly close to Jewett's position, in illustrating something of the semantic richness in Hebrews of πίστις, πιστεύω, and πείθω. Jesus is portrayed as one who ‘trusts’ God and acknowledges his ‘belief’ and ‘confidence’ in God, as well as being himself ‘trustworthy’, ‘firm’, ‘reliable’. The terms are ‘polyvalent’, they display ‘lexical flexibility’, although most often the specific ‘valence’ ‘trustworthy’ is stronger. Thus the overall drift of the argument remains in favour of one or other sense at a time.Footnote 6 Earlier, Richard Hays urged that both distinct ‘objective’ and ‘subjective’ senses of πίστις Χριστοῦ be taken as intended at Gal 3.14. Just at this point Hays came close to proposing the flexibility of language that, it is here to be argued, obtained in general in the world of the NT: uses of ‘faithfulness’ throughout Paul, he insisted, are themselves ‘analogically related’, ‘not [all] identical’. Paul Achtemeier, for one, has remained highly sceptical.Footnote 7
Obviously scholars have long been aware of verbal polyvalence, or polysemity: our dictionaries number on occasion half-a-dozen or more distinguishable senses for lexemes, although, and oddly, ‘polysemous’ itself seems to be taken as unquestionably ‘monosemous’: users of such terms as ‘polyvalence’ assume they can only designate distinct ‘senses’ (‘semes’ or, better, ‘uses’) as discriminated, as enumeratable, in the lexicons.Footnote 8 That usages may merge, unbounded, may flow into one another, implicate one another, seems mostly not to be considered; or if considered, is rejected (‘polysemiophobia’).Footnote 9 One further recent partial exception has been K. F. Ulrichs's cogent syntactical argument that in Paul's use of πίστις Χριστοῦ, both objective and subjective valences evoke one another, complementing each other, as occurs in analogous constructions.Footnote 10 However, he dismisses previous discussions of the wide semantic field of πίστ-words as problematic, and that still without reference to ancient semantics as such.Footnote 11 It seems it can simply be assumed that Paul would have made distinct senses clear: ‘ein genitivus subjectivus eingefürt, erklärt oder jedenfalls gekennzeichnet hätte werden müssen’.Footnote 12
More widely yet, a preference for distinct senses seems generally prevalent in literate circles today, readily found in a standard modern textbook on language. Any ‘crude’ enumeration of senses will certainly be rejected, and more nuanced terms such as ‘facets’ and ‘micro-senses’ will be introduced. But there still seems to remain the ‘modernist’ expectation that the context will make these specific facets clear and distinct.Footnote 13
A problem that as a result confronts this present discussion is the reader's likely tacit assumption that phrases such as ‘the idea of…’ or ‘a meaning of…’ necessarily imply that some precise and definable sense is in question; yet these are the only terms available with which to counter that very assumption. The reader is asked to be patient with ‘scare quotes’ when such terms as ‘meaning’ and ‘idea’ are used. Our use of ‘ambiguous’ is itself also imprecise. Its etymology may suggest (as indicated above) an expression with two distinct senses, or an expression that is simply unclear, in practice polysemous: the latter, however, is the more likely, according to my dictionaries.
The topic mainly to be discussed here is whether our widespread contemporary expectation of and insistence on clarity in the use of words is justified—is at all justified—in our interpretation of texts from the ancient Graeco-Roman world, bearing in mind available contrary indications in ancient commonplaces on words and ideas, on translation, metaphor, and allegory. It is certainly striking that in none of the discussion among NT scholars noted so far does there seem to be any suggestion that a consideration of ancient understanding(s) of semantics might be relevant or can and has been shown not to be. Nor does any such possibility appear to have been considered over the years in a number of monographs on biblical semantics: not in those by James Barr, Anders Nygren, nor Arthur Gibson; nor, indeed, more recently by Anthony Thiselton, nor by others.Footnote 14 Two very recent titles that might seem to suggest concern with these issues also fail to engage with ancient theory.Footnote 15
Scholars today are accustomed to consulting elaborate dictionaries, monolingual and bilingual, where different spellings and senses (or uses) of lexemes are, as just noted, carefully discriminated, even enumerated for us.Footnote 16 Such lexicons only began to appear in Europe in the eighteenth century.Footnote 17 In the ancient Graeco-Roman world individuals compiled word-lists, be it of difficult words in Homer, with suggested current equivalents, or of synonyms and homophones in contemporary use, and speculated on etymologies. But none of even such simple tools as these were in wide circulation.Footnote 18 This already constitutes a very significant difference between authors and hearers/readers then and those of us engaged in literary and kindred scholarship today.
There were occasional attempts to distinguish between the denotations of a general term such as ‘ζῷον’, so a human might be defined as a ‘rational animal’ (ζῷον λογικόν, Aristotle, Chrysippus) or a ‘featherless biped animal’ (ζῷον δίπουν ἄπτερον, ascribed to Plato).Footnote 19 Cicero or Quintilian may explain (with examples) what terms such as ‘paraphrase’ amount to (see further below). They may discuss the limitations presented to a translator by the overlapping usages of Greek and of Latin words.Footnote 20 An individual word in either language is taken to ‘name’ various things or activities, as will be discussed shortly. But there seems to have been no general attempt to distinguish and define (let alone then exclude) possible or apparent connotations of individual words in ordinary discourse—possible senses of ‘human’, for instance.Footnote 21
It is important to recognise that in the ancient Graeco-Roman world only Aristotle earlier seems to have been seriously concerned in general with possible ambiguities that may indeed occur in individual words spoken, heard, written, read. At Topica 1.15.1–30 (106a), for instance, Aristotle distinguishes uses of δικαίως and also of ὑγιεινῶς, with the latter anticipating (or prompting?) my dictionary: ‘denoting health, conducive to health, preserving health’ (compare Poetics 25.22–23, 1461a). But nowhere in his Rhetoric does this sort of analysis recur; nor does it seem to be taken up by later rhetoricians who cite him. In Topica 1.15, Aristotle then notes that ὀξύ can be used as ‘sharp’ is in modern English, in quite different senses in respective contexts of knives and of music. In each context there is now, no suggestion of ambiguity. In fact, he himself at Poetics 25 discusses similar instances in terms of metaphor, the transfer of names (to which, as said, we turn in a moment).Footnote 22
For Aristotle and for later writers ἀμϕιβολία, ambiguitas, is to be mainly discerned in a lexeme such as ΟΥ enunciated with rough or smooth breathing, or a faulty construction, or uncertain breaks between words that leave the performer and/or the hearers unsure, or allow for puns: but ambiguity is not at all said to be found in individual words in contexts seen as normal. In context, different applications are usually assumed to be unmisleading.Footnote 23
Otherwise, it seems that none but the Stoics as a school took deliberate pains to coin or to define their terms so as to obviate particular ambiguities, and they did that for their own distinct philosophical purposes. Roughly speaking, to live rightly you had to have and articulate comprehensively correct impressions of things.Footnote 24 Critics, including the present author, have discerned Stoic influence among early Christian writers; but there is no sign of the latter adopting an extensive Stoic technical vocabulary, essential to any comprehensive assimilation. Otherwise, only Galen, in the second century CE, as an individual philosophically minded physician sought precise clinical terminology.Footnote 25
Philo does suggest that contemporary sophists ‘wear out the ears of any audience they happen to have with disquisitions on minutiae, unravelling phrases that are ambiguous and can bear two meanings (τὰς διπλᾶς καὶ ἀμϕιβόλους λέξεις ἀναπτύσσων)’.Footnote 26 It is to this kind of practice that Mark Given draws attention, explaining how the divergent readings of modern commentators may well represent kinds of double sense that Luke and Paul could have deliberately deployed, with a Socratic transformative irony.Footnote 27 However, this is still to discern distinct senses in words such as δεισιδαιμονεστέρους, ἀγνοοῦντες, ὑπεριδών, and τελός: and that may seem not to go far enough.
For there is no indication that such sophistry (or sophistic anti-sophistry) involved a pervasive attention to individual word usage, nor that such as Philo or Quintilian were drawn in response to pre-empt what they took sophists to be doing for display. Rather did authors (including writers) rely on ordinary words in discursive context to evoke without any ‘disambiguation’ the sort of response sought, as did Stoics themselves for much of the time.Footnote 28 Only in passing does Quintilian tell us ‘in the opinion of certain philosophers, there is not a single word which does not signal many things’ (nullum videatur esse verbum quod non plura significet).Footnote 29 Chrysippus, according to Aulus Gellius, took it that ‘every word is by nature ambiguous, since from the same [word] two or even more things can be understood’ (duo vel plura accipi possunt).Footnote 30
Quintilian gives the suggestion no further attention. And ancient practice suggests that no such general concern arose. Thus when Plato's Socrates discusses topics such as piety, or naming, or knowledge, eros, temperance, manliness, friendship, there seems to be a clear conviction that the discussion is worthwhile, it has a topic, and positions are articulated: but no ‘definition’ of individual terms, their possible nuances, connotations, emerges. A present-day commentator on Plato concludes, ‘To learn the truth we have to go behind words altogether’. ‘Each thing is to be understood through a full, lively awareness of its similarities and differences in relation to other things’.Footnote 31 A first-century bce Platonist offered a very similar observation:
Speaking with multiple voices is characteristic of Plato, and even the subject of the telos is expressed by him in several ways. He uses a variety of expressions because of his lofty eloquence, but he is contributing to a single concordant item of doctrine. That doctrine is that we should live in accordance with virtue. (Τὸ δε πολύϕωνον τοῦ Πλάτωνος. Εἴρηται δὲ καὶ τὰ περὶ τοῦ τέλους αὐτῷ πολλαχῶς. Καὶ τὴν μὲν ποιχιλίαν τῆς ϕράσεως ἔχει διὰ τὸ λόγιον καὶ μεγαλήγορον, εἰς δὲ ταὐτὸ καὶ σύμϕωνον τοῦ δόγματος συντελεῖ. Τοῦτο δ’ἐστὶ τὸ κατ’ ἀρετὴν ζῆν.)Footnote 32
We find much the same in writing in and around the first century. All our authors have read or heard others evoking ‘the idea’ each wishes to evoke in each one's own way: shorter, longer, better illustrated, more elegantly expressed, more or less persuasive, effective: the ‘same’ idea, or near enough for current purposes. But each and every one of these is itself an evocation of, say, ‘the idea’ (‘the “ideal” idea’) of tranquillity, which remains still undefined and undefinable, and with it words such as tranquillitas or εὐθυμία. When Seneca discourses on tranquillity, providence, constancy, anger, clemency; or Plutarch on education, tranquillity, friendship, fortune, virtue and vice, marriage; or Dio of Prusa on kingship, tyranny, virtue, usefulness, or faith (ΠΕΡΙ ΠΙΣΤΕΩΣ), no word and no set of words ‘encapsulates’ let alone itself defines ‘the idea’. It is always outside and beyond the words; and, a fortiori, there is no attempt to discriminate—let alone then to prioritise and exclude—various possible senses, connotations, of the various individual terms deployed, nor does there seem to be on any other topic.Footnote 33 (If exceptions there turn out to be to this present claimed absence of connotative definition and exclusion, they would assuredly seem to be rare.)
There certainly is no sign of any appetite for precise definitions in early Christian writings. When Paul in 1 Corinthians 13 discourses on ἀγάπη he lists a range of ideas the term brings to his mind; however, he does not, for instance, bother to exclude the patriarchal, patronising usage available to the writer of Eph 5.33.Footnote 34 Indeed, we may usefully recall how many centuries it took for Christian intellectuals to agree in stipulating distinct and exclusive uses for terms such as οὐσία and ὑπόστασις; and even then not all were willing to accept the precisions. Precision, and any exclusion of overtones, had to be argued for, if wanted; unargued, they were presumably neither expected nor required. To read ancient authors as though they ‘must’ have shared the concern evinced by some among us for connotative precision risks making a category mistake, a mistake in the genre of verbal articulation deployed.
It is to a more detailed discussion of ancient semantics, and the freedom that they (unawares) may seem to have encouraged that we now turn.
2. Semantic Richness in Ancient Theory and Practice
What follows surveys some ancient authors' discussions of how words are held to work as ‘names’ (ὀνόματα, nomina), ‘names’ that are expected to evoke in hearers' minds shared impressions of people and events and things, and shared ideas, generalities, abstract concepts. Evidence will be offered to show that, with the common acceptance of this model of ‘naming’, there is, as already noted, little call for ‘disambiguation’, but considerable scope for creative imprecision.
In brief, and for example, ‘Diogenes’, ‘the Cynic’, ‘the Sinopean’, and ‘the philosopher’ are all held readily to call to mind the same person (however differently perceived by friend or foe); if others are also called Diogenes, an additional ‘name’ can readily distinguish them, still retaining, undefined, much they have in common (they are Greek males born into Greek culture with its complex and varied talk of Gods and of continuity between Gods and humans…).Footnote 35 And in a similar way a ‘name’ or a short or lengthy sequence of ‘names’ can be expected freely to evoke ‘the same’ but undefined thing or ‘the same’ idea (‘the same “ideal” idea’), be it of friendship, virtue, kingship, variously expounded, in hearers' minds. There can also be a ‘transfer’ (μεταϕορά, translatio) of a name or names from one person or thing or idea to another, expecting a similar evocation without precise definition. In practice, it will be argued, as already indicated, that the corollary is a fluid semantic freedom and richness.Footnote 36
a. ‘Names’ ὀνόματα, Nomina
This understanding of words as ‘names’ arises in all our ancient Greek and Roman discussions of rhetoric, at least in discussions of metaphor, yet seems not to be touched on in the (otherwise very thorough and useful) surveys of ancient rhetoric that have appeared of late.Footnote 37
One of the briefest and most telling accounts available, however, is not in a discussion of rhetoric as such, but in a much-admired example of its practice, the Olympikos of the eclectic Stoic (and occasional Cynic) Dio of Prusa:
The human race has left unuttered and undesignated no single thing that reaches our sense perception, but straightway puts upon what the mind perceives [τῷ νοηθέντι] the unmistakable seal of a name [σϕραγῖδα ὀνόματος], and often several vocal signs for one item [πλείους ϕωνὰς ἑνὸς πράγματος] so that when anyone gives utterance to any one of them, they convey an impression not much less distinct than does the actual matter in question. Very great indeed is the ability and power [ἐξουσία καὶ δύναμις] of humans to indicate [ἐνδείξασθαι] with words whatever occurs [τὸ παραστάν].Footnote 38
In Greek this model for the way words may be taken to work goes back, we gather, to before Plato, and is happily taken up by him and by Aristotle after him, and by their successors.Footnote 39 Plato is able to make a distinction between noun and verb, significantly (re-)using ὄνομα for subject (or noun), with ῥῆμα for predicate (or verb). Aristotle introduces further distinctions (for sentences and connectors), and yet more distinctions follow over the centuries.Footnote 40
There are also discussions, as in Plato's Cratylus, as to whether words originate naturally, perhaps by resemblances of sounds (what we still term onomatapoia); or whether, as most agreed, they are, or are for the most part, arbitrary. There are also discussions of homonyms and homophones, as already noted. However, the unquestioned model remains the same throughout. I quote Jan Pinborg's conclusions:
The semantic conception involved in [Aristotle's] definitions and their context is rather primitive. The written symbols are arbitrary signs of the spoken symbols, which are in turn arbitrary signs of the mental concepts which in turn are natural ‘likenesses’ of the things themselves. This conception presupposes a theory of natural ‘forms’ according to which the forms embodied in the things and giving them their nature is grasped directly by the intellect.Footnote 41
To quote Aristotle himself, as an example of the common view,
Words spoken are symbols or signs of affections or impressions [παθημάτων σύμβολα] in the soul; written words are the signs of words spoken. As in writing, so also is speech not the same for all peoples. But the mental affections themselves, of which these words are primarily signs, are the same for all [ταὐτὰ πᾶσι], as are also the objects of which these affections are representations or likenesses, images, copies [ὁμοιώματα].Footnote 42
Significant is the clear conviction that all humans will share the same mental impressions. As already discussed, Aristotle and his contemporaries and their successors were aware of ambiguity in practice, and others' actions, including ‘inappropriate’ verbal responses, would show whether words had failed to evoke near enough ‘the same’ impression. To ensure success, to allow words their full ‘power to evoke’, speakers (and writers) might use a lot of words, of expansions and paraphrases (and gesture, body-language), to ensure that the intended response was effectively evoked. But ‘meaning’ is in the mind, not in individual words; and quite different sets of words may be expected to evoke ‘the meaning’ intended but not otherwise defined. Failure leads to more discourse; not in most instances, it seems, to individual words' uses being discriminated.
It is worth comparing the passage from Dio with Paul in 1 Cor 14.8–11, for its similar terminology (including awareness of other languages), for its treatment of ambiguity, and for its stress at the end on the ‘power’ (τὴν δύναμιν) of the word. It is the word's ‘power’ to evoke that is at issue, not its ‘meaning’ (pace modern English translators):
…if the trumpet sound is uncertain [ἄδηλον], who will prepare for battle? In the same way, if what you say in tongues produces no clear utterance [εὔσημον λόγον], how can anyone tell what has been spoken [τὸ λαλούμενον]? You will be talking into the air. There happen to be any number of sound-systems [γένη ϕωνῶν] in our world, and nowhere are such lacking. If I do not perceive the force of the sound [τὴν δύναμιν τῆς ϕωνῆς], I shall be a barbarian babbler to the speaker and the speaker to me.
Later, for instance, among the Stoics, ὄνομα comes to be confined in some contexts to proper nouns, denoting named individuals or named things. However, the model remains the same, as one may see, for instance, in Philo:
Who does not know that every language, and Greek especially, abounds in terms [ὀνομάτων], and that the same thought [ταὐτὸν ἐνθύμημα] can be put in many shapes [σχηματίσαι πολλαχῶς] more or less freely [μεταϕράζοντα καὶ παραϕράζοντα], suiting the expression to the occasion?Footnote 43
What for present purposes is most significant is the Stoic insistence that what an assertion asserts, τὸ λεκτόν, is something ‘immaterial, objective, and something which others can grasp’.Footnote 44 A word in common use, and used recognisably, may be hoped, expected to evoke an appropriate notion in any (Greek) hearer's mind.Footnote 45 For most purposes, even among Stoics, there is no need for prior or further definition of individual words, for any restrictions or qualifications to their sense. In practice, speakers in the ancient Mediterranean world, as already noted, learned to use all kinds of amplification and paraphrase in attempting to persuade others to see things nearly enough as the speaker expected. But such amplification, to which we shall return, did not usually include any pruning of semantic riches; if anything, it amounted to further enrichment. Of course, if you were sure you had something quite new and different to say, you would have to define some of your terms, as the Stoics did for their philosophical reflections, and as Galen later did for clinical reasons; but for ordinary purposes both, like everyone else, relied on ordinary language to work as well as they took it for granted it must. Yet, as all the modern authors on ancient semantics cited so far agree, there is in fact no such facility in words to evoke a precise common thought, idea, impression in all users of the language in question. The ‘naming’ model is ‘primitive’ (Pinborg) and unsustainable (Wittgenstein).Footnote 46 ‘Woman’, ‘father’, ‘dog’, ‘freedom’, can be shown to evoke very different ideas and mental images in varying contexts among current speakers of the language. It is worth quoting a later writer, Augustine (criticised by Wittgenstein), for his awareness of this openness:
To be sure, all of us readers try to discern and grasp what the author wished [quod voluit ille quem legimus]. [Yet] what harm is there to me if these words can be understood in different ways, so long as these ways are true? [express divine truth]…even if it is not what the author meant [etiamsi non hoc sensit ille]?… So, when one says, ‘Moses’ thought is mine’, and another, ‘not at all, it's mine’, I think it more faithful [religiosius] to say, ‘Why one more than the other, if both are true? Or if someone sees a third or a fourth, or some truth quite different in these words [si quid omnino aliud verum quispiam in his verbis videt] why may it not be trusted that Moses saw them all [illa omnia vidisse credatur]?’Footnote 47
Moses can be taken to intend to evoke ‘an idea’ or a plurality of ideas, but neither is ‘in’, nor expected to be ‘in’ the words; words only evoke ideas. Neither author nor reader is reprimanded for, nor defended against, any charge of verbal imprecision. There seems to be no thought of engaging in a more precise definition of the terms deployed.
b. Metaphor and Allegory
Ancient talk of metaphor and allegory emphasises the practical imprecision and semantic freedom the ‘naming’ model encouraged. In discussing μεταϕορά, translatio, Quintilian in effect summarises some of the foregoing and prompts further reflection. He explains that μεταϕορά, translatio
adds to the copiousness of language by the interchange of words and by borrowing, and finally succeeds in accomplishing the supremely difficult task of providing a name (nomen) for everything. A noun (nomen [sic]) or verb (verbum) is transferred from the place to which it properly belongs to another where either a proper one (proprium) is lacking or the transferred is better than the proper one. We do this either because it is necessary or because it is more indicative (significantius), or more acceptable (melius).Footnote 48
Obviously, Quintilian, too, took the model of ‘naming’ for granted. Although he could use nomen to distinguish subject from predicate (verbum), ‘metaphor’ was still in both cases the provision of a new ‘name’ (and the examples that follow include proper nouns and verbs). Just as ‘Cynic’ may be substituted for Diogenes to make the reference clear (my example), so crops may be said to thirst. Significantly, another ‘name’ may do as well or better; but no fresh definition, no clarifying precision is called for.
That is not to suggest that ancient authors lacked concern for clarity, lucidity, or effective persuasion and detailed agreement in practice. What is being argued is that the desired clarity was not sought by defining the nuances of even the key terms deployed. ‘Clarity involves the employment of current words, and words bound together’, avers Demetrius.Footnote 49 Only, in agreement with Quintilian, and very significantly: ‘Some things are, however, expressed with greater clearness and precision by means of metaphors’.Footnote 50
This ‘transference of names’ model itself goes back at least to Aristotle (Poetics and Rhetoric), and appears in Cicero, and in the Ad Herrenium.Footnote 51 And it does warrant more reflection than it seems to have received, not least among exegetes of the Judaeo-Christian Scriptures. Janet Martin Soskice, a few years back, argued persuasively that ‘metaphor is that figure of speech whereby we speak about one thing in terms suggestive of another’, and as such it is open and ‘irreducible’ (and many have seemed persuaded).Footnote 52 In a more recent general survey of metaphor Katrin Kohl affirms, ‘Deutlich wird vor allem ihre Kraft, unser Denken, unsere Emotionen, unsere imaginativen Fähigkeiten und unsere Sprache produktiv interagieren zu lassen, sowie auch ihr Potenzial, unsere innersten Gefühlen und abenteuerlichsten Vorstellungen eine Struktur und einen kommunizierbaren Sinn zu verleihen’.Footnote 53 Much recent research in combined psycholinguistics and philosophy of language including pragmatics further confirms such conclusions, and strongly suggests that the use of language with this sort of freedom has been ‘hard wired’ in human brains for many millennia more than the two that separate us from the first century CE.Footnote 54 That is how we for the most part make words work (allow words to work); it is only in some areas of scholarship that verbal precision is sought—or imposed. (On the issue of psycholinguistics, see further the final paragraphs of this essay.)
However, and just because the ancients were convinced that their words could name and rename and so evoke appropriate ideas in others' minds, they will have been particularly free, within the limits of ‘good Greek’ or ‘correct Latin’, to elaborate and innovate expansively.Footnote 55 And we have therefore no warrant for reduction in our interpretation of them. For the model of transferred naming used to explain what we still call ‘metaphor’ is logically symmetrical, reversible. If names are seen as exchangeable, and the transferred one likely as evocative as (or better than) the common one, there is nothing ‘in’ a name that affords precision; precision (if any) lies in the ineffable λεκτόν evoked.
And then, of course, if we, per contra, take it there is no such λεκτόν or ‘form’, there is nothing precise to be named, renamed, evoked, we now have to accept that in practice ancient verbal communication was as open and imprecise as postmodernists have of late argued all language is, but was perhaps even freer because unworried by fears of verbal imprecision.Footnote 56 (Inappropriate some words might be judged to be; that some might be imprecise in normal, non-sophistic use, seems to have been of no general concern.)
Further reinforcement for this argument is afforded by discussions of allegory. After a brief note on catachresis, the use of a term that is quite unexpected, albeit pleasing or otherwise appropriate, Cicero says: ‘When there is a continuous stream of metaphors, a quite distinct style of speech is produced, and so the Greeks give it the term ἀλληγορία. They are right as to the name, but from the point of view of classification Aristotle does better in calling it all metaphors’.Footnote 57
The author of Ad Herrenium uses the Latin term permutatio for allegory, but in much the same way explains, ‘It operates through a comparison when a number of metaphors originating in a similarity in the mode of expression are set together, as follows: “For when dogs act the part of wolves, to what guardians, pray, are we to entrust our herds of cattle?” ’Footnote 58 Quintilian offers a kindred account, but judges the device often tedious, even no better than a riddle.Footnote 59
Stoics and others had long-sought indications in Homer and Hesiod of how the world is. Allegory in Philo works rather differently, but shares the same underlying sense for how words work. So ‘pruning’ (Lev 19.23) can be taken horticulturally, and then as betokening God's generous creative care; but also for ridding ourselves of self-conceit, or doing away with pretence; various possible punctuations suggest yet further possibilities, and Philo interprets the two that most appeal to him: teaching that purifies, or eternal self-evident truth.Footnote 60
Much the same account of the way words work appears to be articulated in 1QHa 9.27–31. Here ‘the mysteries’ to which words refer rest not in them but in the divine mind.Footnote 61 Much the same is implied in Rabbinic midrash.Footnote 62
However, all these examples are drawn from ancient academic discussions and from highly literate and educated authors. It is worth giving the argument a base in common and widespread practice.
c. Paraphrase, Précis, Elaboration, and Amplification
A practical expression of the ‘naming’ model is to be found in the important part that paraphrasing played in learning to read and write. Children learned to retell ‘the same’ story from varying points of view, for various purposes, in varying styles, freely, first orally then in writing, as we are told by Quintilian, and by Theon and others.
Pupils should learn to recount Aesop's fables, the natural successors of the fairy stories of the nursery, in simple and restrained language and subsequently to set down this version in writing with the same simplicity of style: they should begin by analysing each verse, then giving its meaning in different language, and finally proceed to a freer paraphrase in which they will be permitted now to abridge and now to embellish the original so far as this may preserve the poet's sense (tum paraphrasi audacius vertere, qua et breviare quaedam et exornare salvo modo poetae sensu permittitur).Footnote 63
The practice and theory together encourage and warrant the use of metaphor (transferring a name from one idea to another), and so also of allegory. There is no sign of instruction on defining your terms.
To repeat, this is not to suggest that the choice of words was a matter of indifference or simply one of style and elegance. Speakers would have been able to tell from verbal and wider feedback whether ‘the idea’ (simple or complex) intended had been evoked, and had learned to paraphrase with great care. In recent jargon, speakers cared about the ‘pragmatics’ of utterance, their ‘illocutionary force’ and ‘perlocutionary effect(s)’.Footnote 64 This ancient practice was not all that different from that of some postmodernists: you discoursed at length, with paraphrases, examples, illustrations, positive and negative comparisons, but in the hope of attaining an effective resonance and evocation of ‘the idea’ you trusted was somehow ‘out there’ to be evoked. You could make the evocation, the reception of the idea more vivid, more effective, with all sorts of elaboration, just because you could be sure that ‘it’ itself could well be evoked intact.
d. The ‘Intentional Fallacy’?
It might seem that ‘ideas in the mind’ land us back with a version of what was termed last century ‘the intentional fallacy’: the romantic impression that meaning indeed lay inexpressibly in the mind of the author or artist.Footnote 65 The ancient conviction that ‘meaning’ was ‘out there’, in what the Stoics termed τὸ λεκτόν, ready to be evoked in people's minds, is not the same, but it is equally fallacious, as the commentators cited agree. However, two counters are available. The first is an increased (or renewed) current interest in genre, Gattung.Footnote 66 A choice of genre is not a matter of some inner intent. And what it is hoped has been shown in the foregoing is that there is from the ancient Graeco-Roman world no sign of a genre of fine lexical precision, and no sub-genre constituting overtly distinguished connotations for individual terms. To read a first-century Mediterranean document as though it deployed any such sub-genre is a category mistake, a serious anachronism. Secondly, we are not left with translation and interpretation as purely arbitrary, words meaning just anything in a Quinean or Derridean way.Footnote 67 Recent studies in pragmatics show that overt (not hidden) ‘intention’ is inescapably integral to meaning, and ensures that some interpretations are more persuasive than others, even though the pragmatics of ancient communication evidenced in texts rather than as originally performed are, of course, harder to discern than are current ones. However, the pragmatics of ancient rhetoric was discursive. It shows no sign of normally relying, let alone normally insisting on fine distinctions of meaning.
What is ruled out, then, it is here argued, is any hard precision, any clear lines between possible connotations of particular words, the kinds of ‘nice’ distinctions desired in some theological or ideological discourse. In interpreting sympathetically our ancient texts it will, rather and almost inevitably, be a matter of discerning family resemblances among uses of particular lexemes.Footnote 68
For an important illustration of the implications of the foregoing we return to the ‘family’ of terms with which we began: Faith, belief, faithfulness, trust, trustworthiness.
3. Faith, Belief, Faithfulness, Trust, Trustworthiness…
The semantic richness of πίστις and cognate terms (including πείθω) is widely acknowledged, and the wealth of usage is already made clear in the standard dictionaries, whether Liddell and Scott (with Jones and Mackenzie), Bauer (Arndt Gingrich Danker: BAGD), or TDNT (R. Bultmann).Footnote 69
I begin with one specialised context for the use of πίστις, one for which I can find no obvious foregrounded instance in Paul, but an example of which nonetheless helps to set the scene. The lexeme was used by Plato (and by others before him), and again by Aristotle, and then by their rhetorician successors, of endeavours to persuade in court; the conventional but unhelpful translation is ‘proof’.Footnote 70 This usage remains in (UK) English legal jargon in the phrase ‘proofs of evidence’; but the usual modern use of ‘proof’ for a convincingly successful demonstration, ‘evidence sufficing or helping to establish’ as fact (OED) is rather different from an attempt to convince. It is for an attempt to convince, to persuade hearers to trust evidence presented and/or interpreted, that Aristotle uses the term:
Rhetoric may be defined as the ability to discover possible means of persuasion (πιθανόν)… As for ‘persuasions’ (πίστεις), some we do not have to construct [evidence from free or forced testimony, contracts, etc.], others we do, by our own [argumentative] efforts… There are three kinds of these. The first depends on the moral character of the speaker, the second on putting the hearer into a certain frame of mind, the third upon the speech itself, in so far as it demonstrates a conclusion, or at least seems to (διὰ τοῦ δεικνύναι ἢ ϕαίνεσθαι δεικνύναι). The orator persuades by moral character when his speech is delivered in such a manner as to render him worthy of confidence (worthy of trust, ἀξιόπιστον); for we feel confidence (we trust, πιστεύομεν) in a greater degree and more readily in persons of worth.Footnote 71
To believe in someone and so in what they argue is, unsurprisingly, to trust them as trustworthy, be persuaded, convinced that she or he is faithful, and respond trustingly. Thus the passive and middle of πείθω (cf. πιθανόν, above), ‘being persuaded’, ‘being convinced’, is tantamount to believing; indeed the pluperfect passive can be used with the dative or with ϵἰς or ἐπί for trust, rely on (cf. Wis 3.9, quoted below).Footnote 72 The ideas merge into one another.
This complex of belief in, of trust, of trustworthiness, and of persuasion by one trusted, recurs in Ad Herennium, and in Cicero.Footnote 73 Again, in the first century CE, Quintilian notes, ‘All these forms of argument the Greeks name πίστεις, a term that, though properly we may render it fides (warrant), is better translated by probatio (proving, testing)’.Footnote 74 And he, too, insists, that to persuade and gain trust a man must at least appear trustworthy:
Finally, ἦθος in all its forms requires the speaker to be a man of good character and courtesy. It is most important that he should himself possess or be believed to possess those virtues for whose possession it is his duty to commend his client, while the excellence of his own character will make his own pleading all the more trustworthy and will be of service to his cases (Sic proderit plurimum causis, quibus ex sua bonitate faciet fidem).Footnote 75
Philo is in agreement, as is clear in his appraisal of Abraham:
It is stated that he ‘trusted in God’ (ἐπίστευσε τῷ θεῷ). Now that is a little thing if measured in words, but a very great thing if made good in action. For in what else should one trust? High office, fame, honours, abundant wealth, noble birth, senses, strength, bodily beauty?… [All are] precarious. Faith in God, then, is the one sure and infallible good…in him who is the cause of all things and can do all things yet only wills the best… [Such active faith/faithfulness means you] press onward to God by visions of virtue, walking upon a path which is safe and unshaken… God, marvelling at Abraham's faith in him, repaid him with faithfulness (τῆς πρὸς αὐτὸν πίστεως ἀγάμενος…πίστιν ἀντιδίδωσιν αὐτῷ) by confirming with an oath the gifts which he had promised.Footnote 76
It is noteworthy that Liddell and Scott find far fewer instances of ‘believer’ for πιστός than do BAGD. One such in the latter is Wis 3.9, clearly a mistaken choice, but itself significant: ‘Those who have put their trust in him [God] (οἱ πεποιθότες ἐπ’ αὐτῷ) will understand that he is true, and the faithful will attend upon him in love’ (οἱ πιστοὶ ἐν ἀγάπῃ προσμενοῦσιν αὐτῷ). What is at stake in context, especially in Wis 2.10–20, is trust in the trustworthy God, a trust lived out faithfully in love: and to lose the latter strand, faithful living, here picked out by πιστóς, is to misrepresent the text.Footnote 77
Precisely the same, really rather obvious logic is clear in earlier Jewish tradition. The temple furnishers are trusted because they are trustworthy (2 Kgs 12.15). There are no grounds for trust in untrustworthy Pharaoh king of Egypt (2 Kgs 18.21). To proclaim your trust in God is to proclaim his trustworthiness, or at least to attempt to reawaken it (e.g. Pss 25 [24]; 88 [89]).Footnote 78 Although it is possible to imagine someone loyally trusting someone known to be untrustworthy, by and large trust is elicited by trustworthiness, and implies it.Footnote 79
Paul's usage includes at various points at least the range just outlined. He certainly deploys πιστός of the faithfulness, trustworthiness of God (1 Cor 1.9, 10.13; 2 Cor 1.18; 1 Thess 5.24; of Christ, 2 Thess 3.3). Paul hopes that he himself will be found trustworthy by God (1 Cor 4.2–3) just as he has himself found Timothy faithful, and hopes that Timothy will remind the hearers of Paul's own faithful way of life (1 Cor 4.17 with 4.2): and clearly, in context, this faithfulness constitutes a life of lived faith, lived trust. The manner of Paul's lived trust in God as enabler of growth (1 Cor 3.6), ensurer of life (1 Cor 3.22), granter of just commendation (1 Cor 4.5), sustainer in hardship (1 Cor 4.8–13), power in pastoral care (1 Cor 4.20), all display the trustworthiness Paul hopes will ultimately be acknowledged by God (cf. also 1 Cor 7.25). His faith, belief, trust, and his faithfulness and trustworthiness cannot be separated.
When Paul speaks of ‘faithful Abraham’ (Gal 3.9), perhaps H. D. Betz and others are right to prefer ‘Abraham the believer’, despite noting that ‘faithful Abraham’ is a commonplace, for here it clearly is ‘active faith’ that is in focus.Footnote 80 But it is perhaps also worth noting for comparison that in Matthew's version of Jesus' parable, the ‘unfaithful’ slave is the one who distrusted his master (Matt 25.21–26; cf. Luke 19.17–21). Trust and faithful obedience run together. A first-century hearer with any awareness of Abraham in Jewish tradition would most likely understand that faithful Abraham was being commended by Paul, here, for making explicit the trust in God implicit in his (Abraham's) unfailing faithfulness. In faithfully trusting God's promise Abraham displays a crucial aspect of his faithfulness, with still no payment due: thus πιστός as ‘faithful’ in no way contradicts the trust/grace nexus that Paul will clarify later in Romans.
But the gospel parable may prompt consideration of yet another strand in this rich (and unbounded) semantic field. The slave owner entrusted responsibilities to all three slaves: that is, he trusted them. Paul is convinced that God trusts him: that is, God believes in Paul (1 Cor 7.25). God has ‘entrusted’ a ministry to Paul (1 Cor 9.17; 1 Thess 2.4), as he had earlier entrusted his oracles to the Jewish people (Rom 3.2).Footnote 81 But, of course, the same is said in other words much more widely. God entrusts gifts to us, entrusts others' burdens to us, entrusts love for one another to us. God's faith in us is an integral part of his faithfulness, his entrusting to us responsibilities and gifts meant for sharing. (And that may, on reflection, seem rather obvious; it is much easier to trust when trusted, hard in common experience to trust one who does not trust you.)
Clearly, at 1 Cor 7.12–13, we are obliged to translate ἄπιστος as ‘unbeliever’, for here no ‘unfaithful’ outcomes of unbelief are at stake; and perhaps at 2 Cor 6.14 Paul also talks of believer and unbeliever (as commentators seem to prefer), even though it would afford more consistency with the former passage if separation were here ordered for unfaithful behaviour (cf. 1 Cor 5.6), not for the ‘unbelief’ that earlier (1 Cor 7.12–13) was said to constitute no reason for parting. Further, ‘unbelief’ cannot be foremost in mind at Rom 3.3, where human ἀπιστία is contrasted with the πίστις, the faithfulness of God. At Rom 4.1–25, Paul insists that faithful Abraham's faithful trust in God's trustworthiness gratuitously met with ‘justification’ by God. To clear every use he made of πίστις, πιστεύω of overtones of faithfulness, faithful trust lived out in faithful behaviour, would have been nigh on impossible. The terms in ancient use were too rich, the ideas they would evoke too readily elicited together. It is not appropriate for us to impoverish in Paul's writing something his text sees fit to retain in its ordinary richness.
It may well be argued that in Pauline usage at least πίστις ‘names’ specific inner-Christian issues and attitudes not covered in more general usage; but that would not mean that for insiders the term was clear of its common connotations. There is no sign (for instance, in Romans) that earlier hearers had forced Paul to include clear contextual discriminations of this or any other often-recurring set of terms. No such demand figures among the ‘objections’ he tries to forestall.Footnote 82
In his article noted earlier, Barry Matlock examines four passages in which a duplication of πίστις has suggested to some a deliberate attempt by Paul to include the distinct ‘objective’ and distinct ‘subjective’ sense alongside one another: both our faith in Christ and Christ's faith[fullness], respectively. Matlock may well convince others (as he has persuaded the present writer) that it is indeed our ‘trust’ or ‘believing’ that are to the fore in these passages. But Matlock also hopes to have helped to ‘disambiguate’ the usage.Footnote 83 However, what it is hoped has been shown here is that in Paul's world, trust in someone was itself founded in, and displayed and presupposed belief in their trustworthiness (as well as, most likely, their willingness to trust you): faith in Jesus would necessarily imply (unless explicitly denied) at the least a trust in his faithfulness. Ancient expectations of words have them carry much of their semantic baggage with them, whatever part of their range appears in context to be foregrounded; that is, unless some elements of their range have been specifically discarded.
Taking into account ancient understandings of how language works, and noting ancient usage of the key vocabulary, and failing any explicit exclusions from the semantic field of πίστις, the faithfulness of the one trusted is inevitably also there, in the picture, albeit in softer focus.
4. General Exegetical Conclusion
What all this does mean is that we can never justifiably assume that an author in this Greek and Roman culture has intended his or her individual words themselves to contain a precise ‘meaning’, let alone a clear and readily shareable distinct meaning. ‘Names’ are just not expected to function like that. They contain nothing; rather may they summon up, evoke ideas. Ideas of such topics as ‘faith’ or ‘virtue’ or ‘justice’ or ‘freedom’ or ‘law’, it is hoped, are coherent and shared or shareable to some worthwhile degree, but can only be named and more or less elaborately evoked, not in any other way conveyed.Footnote 84 And then no author can be claimed to have used a disambiguated connotation of a lexeme unless or until she or he has made that disambiguation fully explicit. And such disambiguation seems very unlikely in terms of what the ancients said about words and metaphor and translation. If you could trust that a common, even complex idea was already ‘out there’ to be evoked by one among perhaps many common names or sequences of names for it, there was no need to define further the names themselves; indeed, their rich ability in common usage to evoke varied impressions might well be part of, even integral to their power to evoke the particular idea(s) assumed to be on call.
Words, phrases, sentences, paragraphs, spoken and heard, written and read worked then (as they work now), but only by being free to flex and adapt in shared use in life lived together, free to adapt, and not ossified, hardened, made brittle. Sequences of words in our Christian scriptures where we in English (by μεταϕορά, translatio) use ‘faith’, ‘believing’, ‘trust’ words, as with many other such clusters (‘love’, ‘justify’, ‘kingdom’, ‘knowledge’), should be allowed much the same free semantic wealth and varying emphasis as their Greek counterparts enjoyed in the passages we study.
This is, again, not to suggest that the ancients' brains worked differently. It is very unlikely that any major genetic change has occurred over just two millennia. The human language function then was surely just as complex as it is today, as complex as is taken for granted in current debates over psycholinguistics.Footnote 85 Rather would it seem likely that ‘[h]umans behave like jugglers when they use the mental lexicon, in that they have to deal with semantic, syntactic and phonological information at the same time’.Footnote 86 In current discussion there seems to be no suggestion that in the midst of this juggling, between initial inchoate thought and final articulation, there is included a process of checking the range of connotations of each word—or even each leading word—and then preparing to ensure a context (itself also comprehensively scrutinised) that should implicitly preclude every undesired or potentially distracting sense for each term judged important. For sure, we are able to monitor and explicitly to correct our speech just prior to utterance, or soon after; and our writing more readily still, especially with the help of computers.Footnote 87 But there seems to be no empirical evidence for any prior weeding-out of nuances of words as a part of natural language production, today, or, a fortiori, back then. Granted, some scholarly communicators today do monitor some or most of their own and others' written and some at least of their own oral performance and explicitly adjust it to match the norms discerned and announced by lexicographic colleagues—and then try to impose these or other such distinctions on the writings of forbears innocent of any such nicety.Footnote 88 That, it has been argued, is a mistake. Rather do words, then as today, work precisely because their connotations allow them to be used in many settings, and to seem to speakers and hearers to sit comfortably and ‘at home’ where the speaker has settled them, with the context likely foregrounding some connotation(s), but mostly without that speaker feeling any need or desire either to purge or to impoverish the utterance by explicitly exiling others.Footnote 89
