"Daedala Lingua": Crafted Speech in "De Rerum Natura" Author(s): Brooke Holmes Reviewed work(s): Source: The American

Journal of Philology, Vol. 126, No. 4 (Winter, 2005), pp. 527-585 Published by: The Johns Hopkins University Press Stable URL: http://www.jstor.org/stable/3804874 . Accessed: 28/03/2012 12:17
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

The Johns Hopkins University Press is collaborating with JSTOR to digitize, preserve and extend access to The American Journal of Philology.

http://www.jstor.org

DAEDALA CRAFTED SPEECH Brooke

LINGUA: IN DE RERUM NATURA

Holmes

Abstract. This article examines the creation of words in De Rerum Natura through a close reading of two extended passages concerning the problem of where words come from and what they do. The first is the account of speech production, work entrusted to the daedala lingua in Book 4. This physiological process is mimicked at the phylogenic level in the discussion on the origins of language in Book 5, where voice is first shaped by a body responding to the impact of objects, then by utilitas. The adjective daedalus and the intervention of utilitas both signal, I argue, a shift away from an understanding of language as reaction towards an understanding of language as fabrication, a shift with important implications for the relationship of words to the world they represent.

of terms may Lucretius is a poet of reality.1 That juxtaposition call to mind the age-old antithesis between art and science that continues to be a thorn in the side of Lucretian studies, despite the growing consen? sus that De rerum natura might be recuperated as a masterful generic It raises, too, perhaps, the persistent doubts that poetry could synthesis.2 ever have been accommodated as a means of explicating material reality within Epicureanism, whose founder appears from testimonia and extant of deceptive evidence to have been resistant to that genre's repertoire away from strictly literal images and tropes designed to turn language still And, in the sixty years since Paul Friedlander's representation.3 1Unless noted, the translations are my own and the text is Bailey's 1922 OCT. 2 See Boyance 1947; Schrijvers 1970; Thury 1987; Schiesaro 1990; Gale 1994; Volk 2002,72-73. Cf. Ronconi 1963. Synthesis, of course, is a tricky word. All of these scholars in some way or another recognize the novelty of the Lucretian project. 3 This is an ongoing topic of discussion. What appears to have been most objectionable to Epicurus in poetry was, on the one hand, the propagation of fictitious and impious notions of the world and, especially, of the gods in myths (fabulae), and, on the other hand, the use of figurative language, i.e., language deviating from its purely denotative function. Commentators have him rejecting all poetry as 6Ai0piov uuGcgv 8?^eap, the "treacherous trap of muthoi" (Heraclit. Quaestiones Homericae 4.2 = fr. 229 Us.) and denouncing the xd the 'Ouripo-D |icopo^oyr||iaTa, "foolish babbling of Homer" (Plut. Mor. 1087a = fr. 228 Us). In Press ? Journal Philology (2005) of 126 527-585 2005 TheJohns American Hopkins University by

528

brooke

holmes

influential article, "Pattern of Sound and Atomistic Theory in Lucretius" (1941), in which he, seeking to account for the poem's acoustic texture, concluded that the sound-patterns of De rerum natura mimicked pat? terns of atomic reality, it may continue to invite speculation on just how real this poetry is. Friedlander's Lucretius' attempt to understand expert manipulation of Latin verse was built not only on an analysis of Lucretius' style, attention to alliteration, assonance, and rhyme, particularly his meticulous but also on what has become an oft-cited between letters analogy and atoms (elementa),4 five times in the first two (elementa) repeated books of De rerum natura. It is introduced early in the poem, after the statement of the cardinal principle that nothing is created from nothing, to illustrate the principle opposed to such hypothetical chaos, namely, that everything arises from fixed "seeds."This axiom guaranperceptible tees a systematicity to the diversity of the seen world and a limit to what can be created: for "many bodies are common to many things, as we see letters are to words . . ." (1.196-97). that the recurHaving established on what might be cre? rence of common elements constraints imposes of the analogy, ated, Lucretius, in a second, more elaborate deployment shifts his emphasis to the manifest diversity of nature, which he ascribes to variations in the unions, positions, and movements of the atoms. He to letters, which also combine in various ways to form lines and points words that differ in sound and meaning:

the last decades, scholars have added considerable nuance to this picture: see especially Asmis 1995a, 16-21, who demonstrates the ambiguity of Epicurus' position in these sources; see also Gale 1994,6-18; Wigodsky 1995; Obbink 1996. On the second objection regarding figurative language, see Ep. Hdt. 37-38, 72-73; On Nature fr. 12 col. iii 6-12 (Sedley); D.L. 10.13. See also De Lacy 1939; Boyance 1947; Classen 1968; Long 1971,123-25; Sedley 1973, 18-23; Dionigi 1988,70-73. These qualities primarily affect poetry's success as a medium of philosophical exposition, and Epicurus may not have had anything against deriving pleasure from it, provided one's worldview had already been firmly established by philosophi? cal instruction, as Asmis suggests (1995a, 21). Such an understanding of poetry accords with what we find in the fragments of Philodemus. Despite being an adroit epigrammatist, Philodemus did not find any philosophical utility in poetry. He seems to have viewed it as a possibly pleasant diversion not without the potential to do harm; see Asmis 1995a, 26-33, for an overview of his position. 4 There is a long history of this double meaning, beginning with the use of the Greek word stoicheion to mean both "first-element" and a letter of the alphabet; similarly with the Latin elementum. The double entendre probably originated with the Greek materialists; see Arist. Metaph. 985bl0-21; De gen. et corr. 315a35-bl5. See also Diels 1899,9-14; Snyder 1980, ch. 2; Ferguson 1987; Dionigi 1988, 20-38; Porter 1989, 157-78; Armstrong 1995, 210-13.

CRAFTED SPEECH IN DE RERUM NATURA namque eadem caelum mare terras flumina solem constituunt, eadem fruges arbusta animantis, verum aliis alioque modo commixta moventur. quin etiam passim nostris in versibus ipsis multa elementa vides multis communia verbis, eum tamen inter se versus ac verba necessest confiteare et re et sonitu distare sonanti. tantum elementa queunt permutato ordine solo. at rerum quae sunt primordia, plura adhibere possunt unde queant variae res quaeque creari.

529

(1.820-29)

For the same [sc. elements] form sky, sea, lands, rivers, sun; the same create, too, crops, trees, living beings, but only when combined with different elements and moving in different ways. And, moreover, in every part of my verses you see many letters common to many words, although you must grant, nevertheless, that the verses and words differ from one another both in their sense and in the sound of their sounding. So much can the letters do with only a change in order. But those elements that are the firstbeginnings of things can employ even more means by which all different things may be created. lines later, he gives the example of ignis (fire) and lignum (firedistinct letters that produce, nevertheless, wood), words with common distinct objects, just as the objects sounds (distinctae voces) denoting which differ in their positions, share elements themselves movements, with other atoms in the compound.5 The word, then, is and interactions of a perceived the visible artifact that reveals the underlying composition for both the differ? It clarifies the similarity that is responsible object.6 and the ence heard in the spoken word and seen in the thing denoted Some of those words and objects. recurrence systematic The analogy returns with yet another twist midway through Book 2. of the equivais often seen as a simple reiteration While this incarnation atoms and letters, Lucretius lence between is, in fact, making a novel the same which recurs in essentially point. In the version at 1.820-29, the critical role of position the analogy illustrates form at 2.1013-22, For the same letters signify (significant, in word formation. (positura)

5 "iamne vides igitur, paulo quod diximus ante, / permagni referre eadem primordia / cum quibus et quali positura contineantur / et quos inter se dent motus accipiantque, / saepe atque eadem paulo inter se mutata creare / ignis et lignum? quo pacto verba quoque ipsa / inter se paulo mutatis sunt elementis, / cum ligna atque ignis distincta voce notemus" (1.907-14). 6 On analogy in early Greek science, see Lloyd 1966; for Lucretius, Schiesaro 1990.

530

BROOKE HOLMES

on how they are arranged. Ordo de2.1016) different things, depending as well. In each and 1.906-14 in words at 1.820-29 termines difference as an epiphenomenal case, the diversity of the sensible world is represented and reordering.7 Yet the analogy at 2.688effect of atomic redistribution 99 is part of an extended explanation, beginning at 2.333, of the variety of and so its aim is to demonthe atoms themselves, by shapes possessed strate that, while there are many letters common to the poem's words and verses, they are, nevertheless (tamen), formed from different elements of his prior is clearly conscious Lucretius ex aliis . . . elementis). (alia after having made clear what is for he elicits acquiescence only emphasis, not to be admitted this time: "[you must confess] not that but a few letters run through them in common, nor that two of them are made of letters all but that they are not all alike the same one with another."8 indeed, yielding difference Language proves a rich source of illustration a closed system and essentially, while maintaining both epiphenomenally If it is still unclear to what extent letters them? of possible formations. to the atoms whose nature they betray and whose selves are indebted the same, mimic?are their own collocations building they themselves compounds blocks of a real world? what kind of object is a sound? a word? a more pressing question might be: what exactly are these ele? poem??a to us not only the written For Lucretius presents ments of language? marks on the page but also the sounds and the things (res) they produce. and repetiCritical allegiance to one of these systems of difference of the anal? tion has produced a spectrum of competing interpretations himself was for Lucretian poetics. Friedlander ogy and its implications with the acoustic texture of the poem. For example, the repconcerned etition of sounds in religione and caeli regionibus "can hardly be a mere affair of sounds. The sounds express a reality, the fact that religion de7 Of the three Aristotelian terms (thesis, taxis, schema), the pairing thesis and taxis predominates in the analogy with language, as the evidence of Porter 1989,149-78, shows. 8 "quin etiam passim nostris in versibus ipsis / multa elementa vides multis communia verbis, / eum tamen inter se versus ac verba necesse est / confiteare alia ex aliis constare elementis; / non quo multa parum communis littera currat / aut nulla inter se duo sint ex omnibus isdem, / sed quia non vulgo paria omnibus omnia constant. / sic aliis in rebus item communia multa / multarum rerum eum sint primordia, verum / dissimili tamen inter se consistere summa / possunt; ut merito ex aliis constare feratur / humanum genus et fruges arbustaque laeta" (2.688-99). The contrast (see the repetition of tamen at 690 and 697) between the recurrence of common elements among ostensibly different things and the differences between these common elements (each has a dissimilis summa), which result in differences of compound objects that are not reducible to order and arrangement, has been overlooked, although see Volk 2002, 103. Dalzell does note that Lucretius is making a different point here but finds the argument inchoate and incoherent (1987, 22-23).

CRAFTED SPEECH IN DE RERUM NATURA

531

rives from a heavenly region" (1941,19). Such play exploited a theory of language, he argued, that related sounds to the nature of the objects they with similar words bearing witness to relationships of affinity denoted, between the objects designated, and the poem itself "repeat[ing] the creative work of language on a different level" (1941,29). Whether or not Epicurean language theory explicitly provides for this kind of relation? ship between words and things will be taken up below. As for Friedlander's Dalzell (1987) has observed Alexander that his exclusive "atomology," attention to resemblance Lucretius' point about the emergence neglects of qualitatively different compound objects as a result of slight changes to the arrangement of their component elements. But Dalzell himself can only see the production of epiphenomenal difference, an illustration of how difficult it is to hold the competing vectors of the analogy mentally. In the last thirty years, a number of scholars have extended, fineto Friedlander's Where he sought to tuned, and responded arguments.9 break down the analogical barrier between the elements of language and those responsible for sensible reality by positing a natural (i.e., nonbetween the utterance and its referent, others conventional) relationship have either preserved and complicated the analogy or, conversely, sought other strategies for integrating language into a world where bodies act and are acted upon. Recently, the non-conventional position has been David Armstrong in light of ongoing work on the developed by (1995) a poetics committed in its moments of positive poetics of Philodemus, to the radical formal integrity of the poem, that is, the insepaexposition Just as the rability of its verbal style (lexis) and its thought (dianoia). transposition (metathesis) of atoms in the Lucretian analogy yields a completely new object, any change to the poem, even of a single letter, should create a different literary artifact, Armstrong argues.10 The analogical

9 See especially Snyder 1980; Ferguson 1987. Deutsch 1978, which first appeared as a Bryn Mawr dissertation in 1939, initiated the systematic inquiry into patterns of sound in Lucretius. Also, Ivano Dionigi has detailed how the language used by Lucretius to speak of atoms borrows a number of technical terms from the grammarians.Analyzing the ways in which Lucretius maximizes rhetorical and stylistic strategies (e.g., anaphora, leonine rhymes, alliteration) to create a tautly structured verbal artifact capable of reproducing "la struttura del reale," he goes further than Friedlander and Snyder to claim that words and things are intertwined in such a way as to reveal "una funzione ermeneutica prima ancora che pedagogica ed estetica" (1988,109). For an overview of "atomological" criticism, see also Volk 2002,100-105. 10 Armstrong 1995, 228-32; see also Dionigi 1988, 33. Armstrong emphasizes the of written letters. However, the symbiosis of letters and phonemes should be acknowlplay edged, for the poem would not have been read silently, as Lucretius' references to the

532 barrier

BROOKE HOLMES

remains firmly in place, since for Philodemus the play of written letters is epideictic show, producing a purely intellectual pleasure that is derived from watching a miniature kosmos take shape. The effect, that is, is devoid of moral utility. Yet this barrier may be more porous for Lucretius. It is hard to conclude from his metapoetic statements that his sole aim is simile (1.921intellectual The basic tension of the honeyed-cup pleasure. 50 = 4.1-25) between pleasure and philosophical instruction undermines such an interpretation of Lucretius' and complicates any identification thinks he is doing with those of Philodemus. Lucretius poetics clearly

with words beyond bellelettristic something play. A fascination with the Lucretian has marked the performative route taken by other heirs to atomological criticism. Rather than elaborate Friedlander's about the relationship of sounds to the hypotheses nature of things, they have focused, rather, on the status of the poem as a physical object. E. M. Thury (1987) has argued that the poem's visual of the rerum natura, taken as a whole, obeys the same representation laws as any atomic image (simulacrum) apprehended by a viewer: Lucretius' is to provoke a "clear view" of the kosmos. Alessandro goal Schiesaro (1994; cf. 1990,21-30) too has focused on the images triggered the written words, claiming that uthe poem creates a wealth of mate? by its status as sheer medium" (1994, rial objects in our mind, transcending he speaks of the poem as a material body, this is, he 88). Although for concedes, "shorthand" for saying that the poem is a set of instructions a material body, that is, a series of simulacra, in our mind. Yet creating with this gesture, the nature of the words themselves fades before the what Schiesaro clarity of the image, making it difficult to understand means when, a page later, he refers to a string of words as a string of material objects, not like atoms, but simply atoms. As Katharina Volk "shorthand" ends up taking (2002,103-4) recently observed, Schiesaro's the place of an argument that might explain how the poem participates in the same order of reality as the simulacra it produces. This is not a for Volk, since she sees the relationship letters and between problem atoms as a strictly illustrative analogy. While I share her misgivings about Schiesaro's approach, I believe that Thury and Schiesaro, in linking words and simulacra and in seeing this linkage as fundamental to the references

"ears" of his reader mark (e.g., 1.417; 2.1024; 4.912; 5.100). As Friedlander insisted, the poem was not only a written artifact but a heard one. Porter 1989 discusses at length the implications of this analogy for metathesis more generally in atomist poetics, which he sees upheld not by Philodemus but by some of Philodemus' opponents in On Poems, i.e., Crates and the kritikoi; see also Porter 1996.

CRAFTED SPEECH IN DE RERUM NATURA

533

Lucretius makes to his own practice, are essentially correct. It is certainly a position encouraged Lucretius' own rhetoric of lucidity. Yet there is by a risk in embracing too readily this rhetoric, which emphasizes revelation even as the poet refuses to render transparent of sound the medium for producing that revelation responsible qua epiphenomenon. In what follows, I would like to reexamine the status of the poem as a created object by looking at what Lucretius has to say about the creation of speech in general. While the process of how words communicate things is only hinted at?meaning to this that some background must be supplied from what little we know of Epicurean sequestion mantics?I believe there is much to be gained for our understanding of the relationship between language, reality, and artifice in De rerum natura from a close reading of two extended passages, which, each in their own way, attempt to account for where words come from and what they do. The first passage is drawn from the explanation of the production and of speech in Book 4, where the fashioning of voice into perception words is entrusted to the daedala lingua. This physiological intelligible is mimicked at the phylogenic level in the discussion of the process of language in Book 5, where raw voice is shaped into discrete origins sounds first by a body reacting to the sensory impact of objects, then by utilitas. Both accounts an understanding of language, and encourage as a deictic exercise designed to indicate objects in particularly naming, the world. Thus, both support Lucretius' understanding of his own work as a process of rendering visible. This investment in the transparency of the verbal artifact belies the complexities involved in translating experi? ence into sound and sound into vision, a sequence of events that, in the earliest stages of language, has its systematicity secured by its subordination to natural law. This sense of automatism dominates both of the Lucretian passages, even as we begin to infer the presence of an intentional subject. I will argue that it is in the ambiguous stage of utilitas that the lingua becomes both natural daedala, a curious epithet describing and human creation in De rerum natura. How utilitas comes to bestow this title on the tongue raises questions about the nature of speech qua fabrication, the autonomy of language vis-a-vis sensation, and its prospects for representing the real, visible and invisible.Thus, although Democritus' statement that the "architect" Homer built a kosmos out of all kinds of words with Lucretius' desire to "build a poem" (DK68 A21) resonates condere, 5.1-2) worthy of the master, I would like to begin with (carmen the world fashioned each time we open our mouths. In addressing what takes place at this moment, we might gain a better understanding of Lucretius' sensitivity to the uses and abuses of language.

534 I. THE

BROOKE HOLMES IMAGE OF A WORD

Hasce igitur penitus voces eum corpore nostro exprimimus rectoque foras emittimus ore, mobilis articulat verborum daedala11 lingua formaturaque labrorum pro parte figurat.

(4.549-52)

Then when we express these voices from the depths of our body and send them forth directly from the mouth, the nimble tongue, fashioner of words, joints them and the shaping of the lips in turn gives them form. Hearing is not the first but the second sense in this regard is telling. The and its belatedness and associated topics, begins with an perception the introduction of the simulacrum,12 a concept explained by Lucretius, fourth book, devoted to of vision and exposition inherited from the early the terms in which Lucretius Greek Atomists that is crucial to establishing senses but mental processes as will explain not only the other non-tactile which are "stripped just like a well. Objects steadily shed simulacra, from the surface of a body and float this way and that through membrane films strike the eye, we "see" their the air."13 When these image-bearing of the simulacra source. Of interest at the outset is that the relationship For before any talk of as a kind of afterthought. to sight is described seeing, the simulacra appear mid flight, without any particular destination.The object emitting them does not intend to be seen, and thousands without flit aimlessly of simulacra, (volitant) generated unremittingly, ever striking an eye. It is this world, overpopulated by object-avatars? for thought later in the of the filmier simulacra responsible anticipations rises up before us in the beginning of Book 4. book?that After a discussion of the simulacrum's marvelous velocity and a lacuna of undetermined length, the simulacra finally impact an eye and is cause an imago to appear, at which point a material relationship established between the perceiving subject and the perceived object, the space between them bridged by the continual stream of simulacra from the object (cf. Ep. Hdt. 50). Couched in these terms, it is clear that sight is

11 OQ have nervorumdaedala.Editors have unanimouslyaccepted Lambinus'verborum. 12Lucretius rarely transliterates and so drops the Greek eidolon. Simulacrum is his primary translation, but he also often uses imago as well as efftgies and flgura, terms that still emphasize the eidolon's iconic quality; see Sedley 1998,39-42. Cf. Epic. Ep. Hdt. 46-51; Diog. Oen. fr. 9 (Smith). 13 "quae, quasi membranae summo de corpore rerum / dereptae, volitant ultroque citroque per auras" (4.35-36). See also 4.54-64; Ep. Hdt. 46.

CRAFTED SPEECH IN DE RERUM NATURA

535

modeled on a tactile relationship. Indeed, the appeal to touch constitutes the last recourse to the reader's experience to confirm the veracity of the simulacrum before the detailed discussion of vision:14 Praeterea quoniam manibus tractata figura in tenebris quaedam cognoscitur esse eadem quae cernitur in luce et claro candore, necessest consimili causa tactum visumque moveri. nunc igitur si quadratum temptamus et id nos commovet in tenebris, in luci quae poterit res accidere ad speciem quadrata, nisi eius imago? esse in imaginibus quapropter causa videtur cernundi neque posse sine his res ulla videri.

(4.230-38)

Moreover, since a shape handled in the dark is known in some sense to be the same as that seen in light, in the clear brightness, it must be that touch and sight are moved by a similar cause. Therefore, if we touch a square thing and it affects us in the dark, what square thing could fall under our gaze in light if not its image? On account of this it is obvious that the cause of seeing is in the images, and that nothing can be seen without them. Note that touch only moves us in the dark. In luci, the square ceases to with us directly, falling, rather, under our vision (ad speciem) and to us what we had identified in the dark through touch, as representing touch were no longer an option. Touch is unmediated contact, but though it is the derivative sense that takes over as soon as there is light. Touch the truth of vision even as it is superseded legitimates by it, and the is surpassed by its imago; it is sight above all that gives perception object of the things themselves, res ipsae (4.258).15 This is consistent with the of vision in Lucretius as well as in Epicureanism and Greek privileging as a whole. However, the tradeoff is worth noting, for it philosophy other exchanges in which contact with the world gives way to anticipates interact

14Cf. 1.304;2.408-9; Cic. ND 1.49. On the importance of touch, see Bailey 1928,4046; Schrijvers 1970, 88-91; Rosenmeyer 1996,142-43. 15See Glidden 1979, who argues that Lucretius believes perception is perception of objects rather than sensory impressions. This is true insofar as we take vision as a model. Yet other forms of sensory perception are strongly associated with pleasure and pain, and sensation is, in fact, polarized in order to correspond to these two extremes (see Graver 1990, 98, for example, on taste). It is thus more difficult to postulate an absolute divide between perception and sensation (aisthesis and pathe for Lucretius are both translated by sensus, as Glidden details).Tellingly, the act of hearing words concentrates on the percep? tion of an external object rather than on the sensation they provoke, as I show below.

536 the contemplation

BROOKE HOLMES of it and more distortion distant things, albeit at the risk of

introducing greater Given that vision is ultimately established on analogy with touch, it is not surprising that problems arise because such contact is not immediattributes all perceptual error?that is, the failure to propwhat is?to the viewer's of sense-data erly identify misinterpretation either disrupted in the course from the object to the viewer or otherwise skewed by the conditions of viewing. Thus, air wears down the sharp of a square tower, causing the simulacra to arrive rounded, their angles In such cases, cf. Sext. Emp. Math. 7.206-10). angles blunted (4.353-63; the loss of spatial immediacy is gained by the (temporal immediacy visual simulacrum's awesome speed) mitigates the power of the simulacra to represent the distant object with total fidelity, although the mind may of make adjustments based on what it knows about the conditions viewing.16 In his treatment body capable and a percipient: of hearing, Lucretius retains of establishing a material relationship the idea of a micro between an object ate. Lucretius

and error.

principio auditur sonus et vox omnis, in auris insinuata suo pepulere ubi corpore sensum. corpoream [vocem add. Lachmann] quoque enim constare fatendumst et sonitum, quoniam possunt impellere sensus. (4.524-27) To start with, every sound and voice is heard when, having crept into the ears, it impacts the sense with its body. For it must be granted that voice and sound are also corporeal, since they can strike our senses. The argument benefits from the proximity of the proofs for the existence of simulacra, which had explained in prothe role of such micro-bodies Yet further evidence is deemed to prove voking perception. necessary the corporeality of sound.17 Turning to the sound-producing body, Lucre16On perception and perceptual error, see further below pp. 546-47. 17 The corporeality of voice was a controversial issue in ancient philosophy; see Koenen 1999a, 439^0; id. 1999b, 23; Biville 2001, 28-30. Sonus/sonitus generally refers to all auditory stimuli, while vox is restricted to sound produced by living beings, including animals; vox articulated will become the word. See Arist. Hist. an. 535a28-536b23, where he distinguishes voice, sound and speech (phone,psophos, dialektos), of which only humans produce the latter; Arist. De an. 420b6-421a7; Int. 16a27-29. For later attempts to distin? guish human speech from that of animals (including trickier cases such as parrots), see Gal. In Hipp. Off. (18b.649 Kiihn - SVF 2.135); Sext. Emp. Math. 8.275 (= SVF 2.223); I owe these references to David Armstrong. The Stoics, according to Diogenes Laertius (7.55-56),

CRAFTED SPEECH IN DE RERUM NATURA tius

537

of the "first-beginnings of voices" argues, first, that the passage the throat can hurt: pain proves that voices (primordia vocum) through and words are made of corporeal elements.18 He then goes on to point out that a speaker becomes exhausted, having given away a part of her

sound is thus felt in a way body (amittit de corpore partem). Producing that being seen is not: it is a kind of work.19 The rhetorical move here makes perfect sense given Lucretius' tendency to appeal to the experi? ence of the reader whenever topossible, yet the decisive orientation wards the rest of his speech at this point will reverberate throughout discussion of hearing.20 Although lines 4.542-48 seek to account for the quality (rough or smooth) of sound, which is determined by the shape of the primordia,21 lines 4.549-614 revert to the direction taken in the proof of sound's corporeality and treat speech alone.22 Thus, when Lucretius

believed that the phone of an animal was air that had been struck by an impulse (&f|pi>7i6 while opufjg7t?7ttaTy|jivo<;), human speech is articulated and sent forth from thought (evccpGpcx; kou oc7io On 8iavoiac; ?K7i?U7io|jivr|). the notion of horme in Stoic thought, see Stob. 2.86.17 (= SVF III.169). Philodemus, too, binds articulation to thought; see Blank 1995,183-84. 18Cf. 2.436; D.L. 10.32 5e kou xo (i)cp?GTnKExo te opav r|uxx<; cckoueiv,coorcep dXydv). 19For the proofs, see 4.528-41. On Lucretius' awareness of how the body produces voice, see Gourevitch 1997; Koenen 1999a, 441-42; id. 1999b, 24-25; Biville 2001. 20The central two arguments on the corporeality of sound are drawn from speech. Thus, Bailey is being disingenuous when he says that "Hitherto Lucr. has argued that the hearing of sounds is due to the emission and reception in the ear of emitted particles. Now he deals with the special case of speech, in which there is added to the sensation (aisthesis) a perception (epaisthesis) of significant words" (1947, ad IV.549-94) and references Ep. Hdt. 52. Moreover, Bailey's interpretation of epaisthesis as the perception of words is controversial. In Epicurus, it pertains to all perception, without involving logos: see Lee 1978, 37-38; Asmis 1984 113, n. 21; 162-63. 21The position of the passage is much disputed and line 545 hopelessly corrupt. For a detailed discussion of various editors' conjectures, see Koenen 1999a, 452-53; id. 1999b, 27-32. For Epicurus, the qualities of color, smell, taste, and heat are properties of the object, whereas Democritus attributed them to our subjective sensory experience; see DK68 A135, B9; and Bailey 1928,168-74; Furley 1993. This is treated at length by Lucretius at 2.381-477. On the specificity of the quality of sound, see 2.408-13; 4.542-48. 22See Schrijvers 1974,351, n. 40, and Koenen 1999a, 455. Both argue that the way in which Lucretius proceeds is in keeping with the traditional arrangement of topics in ancient philosophical and medical texts, such as Hp. Carn. 15-18 (Littre 8.603); id. Vict.2.61 (Littre 6.574-75); id., Morb. Sacr. 16 (Littre 6.390); Arist. Hist. an. 532b29-536b23; De an. 418a27-24al6; and ps-Gal. Def. Med. 41-44 (19.358-59 Kiihn). In the majority of these cases, while it is true that speech is explained in the context of the other senses, hearing and speaking qualify as two different things, with the discussion of voice and speech usually following the material on hearing, as at Arist. De an. 419b4-21a7.The treatment of hearing usually deals with the body of the listener, while that of the voice addresses the speechproducing body and thus does not address hearing words.

538 describes

BROOKE HOLMES

the flight of the sound particle towards an ear (4.553), we are only dealing with an articulated word. Other forms of hearing disappear, and with them the attention to the nature of the sound itself, rough or smooth. The dynamics and a of the relationship between a speaker listener take center stage, a relationship to which Epicurus himself does not give special attention in our extant texts.23 of voice ex? Lucretius' decision to pass over the first-beginnings from inside the body, whose qualities of roughness or smoothpressed ness are invariable, in favor of speech raises the question of the body's in the production active involvement of sound. It is difficult to imagine an object here simply emitting sounds, as an object sheds simulacra and unconsciously, because the topic has shifted continually, automatically, towards bodies whose production of sensory data is episodic, provoked and defined by these bodies' undertaken, by a stimulus or deliberately to vary the kinds of sounds that are produced.24This capacity for capacity of which is also the condition for the intentional variation, manipulation in will prove crucial to the development of language described voice, Book ments 5. Moreover, in articulated Lucretius' words, of sound particles as ele? presentation a subtle shift that suggests a breach in the

23In his reconstruction of the relationship between Epicurus' On Nature and De rerum natura, Sedley 1998 aligns the discussion of the other senses in the former, listed as section xv (following section xiv, "vision, visualization, truth and falsity") in his Chart I (133), with 4.522-721 of the latter. He believes that this lost section of On Nature would have corresponded to Ep. Hdt. 52-53 where the nature of auditory particles (onkoi) is treated. Speaking there is classed with making any kind of akoustikon pathos. Words are not treated with any specificity, and so the manipulations of particles by the mouth are not mentioned. Rather, sound originates as a blow (plege)?its cause is not noted?that occasions a "squeezing out (ekthlipsis) of certain bodies"; these create a current of breath (pneumatodes rheuma). It is the specific nature of this rheuma that produces the appropri? ate perception in the hearer; see Lee 1978,31. See below, p. 541-42, for the implications of this account for Lucretius' own description of sound transmission. Koenen (1997,167), of course, is right to point out that the absence of some elements of Lucretius' own explanation of sense perception from the text of Letter to Herodotus does not mean that those elements did not appear in the discussion of the senses in Books 3 and 4 of On Nature in Sedley's reconstruction. A discussion of articulated voice does appear in the fragmentary treatise on the sensations attributed to Philodemus; see PHerc 19/698 col. xxvi-xxvii. It is telling that only articulated voice (engrammatosphone)/articu\ation (arthrosis) has aschema here, suggesting that Lucretius' treatment probably draws on something from Epicurus or from the post-Epicurean tradition. See below, n. 26. 24Koenen distinguishes between "automatical" and "non-automatical" emissions (1997; 1999a, 436; 1999b, 21). The former are involved in vision and olfaction while the latter require some sort of cause, whether it be a person squeezing voces out of her body or the aggregates of flavor-constituting atoms extracted through mastication.

CRAFTED SPEECH IN DE RERUM NATURA letter-atom

539

analogy has occurred at the level of the spoken word, also how we are to understand For voces, their referentiality. complicates once fashioned into words, do not, at first glance, seem to communicate about the body that produced them (or rather, contributed anything itself to them), nor is that body-gwa-macrophysical object the thing about which the listener is seeking to gain information.25 What words bear, is information about the world. We will see that this act of making rather, reference does implicate the sound-producing body, but the fact that it also go beyond that body encourages us to note the difference at may this point between the cries of animals, where variation expresses states of the body?such as pleasure and pain (5.1056-90)?and human lan?

of things in the world. It is guage, where variation reflects the multiplicity the failure of the sounds of animals to go beyond the body and precisely indicate the external world that leads Lucretius to define them as mute cf. Arist. Hist. an. 536a-536b).Thus, the problem of distortion (5.1059,1088; in the simulacrum's transit in the discussion of hearing does not endanger the sound particle's fidelity to the body emitting it but rather its to another "original." What is this, if not a property of the body? fidelity If hearing is understood as hearing words, auditory "simulacra" are in a strange position. For while continuing to assimilate the sound particle to the simulacrum via its corporeality and mobility, Lucretius has embarked on a treatment of a different kind of mimesis, albeit one linked to vision. The simulacrum is iconic. Its ability to conserve closely the form of the object, its thesis and its taxis for Epicurus, is explained by the fact that it is thrown off from the body's surface (summo de corpore). Thus, in its flight from the body, it is not impeded by anything (4.59-64), unlike smell, smoke, and heat, which arise from the depths of a body (ex alto, intrinsecus, 4.90-91) and are "torn up" as they exit. Any distortion of the simulacrum'^ form results from its time in the air, as we have seen.

25For physiognomists, physicians, and the orators, a vox communicated a wealth of information about the body that produced it; Biville 1997 gathers evidence from Latin literature on the vox that signifies without or in excess of the verbum. The relationship between a sound and the sound-producing body (that is, the sound of an oboe, the sound of Madonna's voice) is important to the Epicurean account; see Lee 1978,34. However, in the Lucretian description of meaningful sound, while a word may retain the property of roughness, this is irrelevant to what it represents. The various sounds animals have at their disposal to express emotion in Book 5 are also available to humans (e.g., 5.996, the expression of pain), just as singers may imitate birds in "leading their voices in various ways" (ducere multimodis voces, 5.1406). However Lucretius concentrates on speech as the natu? ral outgrowth of humans' ability to vary the sounds they produce in the interest of referring to the world rather than simply themselves.

540

BROOKE HOLMES

The spoken vox also has a form that must be preserved (4.556).26 But what kind of form is this? More generally, what does it mean for it to be created rather than spontaneously shed? The word is voice that the tongue has divided up (articulat, 4.551) into parts.27 Each is then fashioned (figurat, 4.552) into a determinate work that falls to the nimble tongue, "artificer of words," and the shape, forming of the lips (mobilis . . . daedala verborum lingua formaturaque the word as a "jointlabrorum). Articulatim, at 4.555, further emphasizes of fashioned sound particles into a specific linear ordering. This ing" anal? ordering is crucial for, as Lucretius makes clear in the letter-atom it impacts how something sounds and, thus, which object will be ogy, denoted. Each sound must be first articulated by the speaker, then picked out (discerni) by the listener before the word can be said to have been francommunicated.28 Verbal form, it seems, is a compound, successfully gible, created thing, with transit (4.556). This figura in its own figura, which must be preserved is not exactly the copy of a word, since else-

26 servat enim formaturam servatquefiguram. Notice the repetition of the verb servare here and at 4.69, where it is used with simulacra (formai servare figuram). Cf. 4.87,100; Ep. Hdt. 48: Kcdydp pevciq anb xcovacoumcovxov emnoXriq xf|v eni xov guvex^q . . . acp^o-oaa ax?p8uvio\) 0eoiv Kai xd^iv xcovdxoucov?7tinoXvv xpovov. . . . See also PHerc 19/698 xxvi (Monet), where form (schema) is said to be the object of discernment (krima) available to x[p]od<;I r\ x[6] ?vypa(i|j,dxo\)I every sense. The author then goes on to speak of xo xf|[<;] (pco[vfi(; a%r\\ia_The analogy between color and articulation is continued in the next lines. The author apparently is comparing the relationship between the form of color (6 xov and x[p]cb|iaxo<; oxnuxxxiafioc;) the quality of color to the relationship between articulation and the quality of the voice (npbqxr|v xfjq(pcovfjc; He (fj dp9pcooi<;) rcoioxrixa). concludes by xcvuxo o\>k axril|iaxo<; saying that neither color nor voice is a stranger to form (kou kocxoc dM,oxp[(av] I eutojiev dv ?Kaxe!paveiv[a]i [x]covaia0ril[a?cov . . .). 27 Articulare translates the Greek verbs dpOpo^v/SiapOpovv, technical terms used to describe phonetic articulation. See Hp. Carn. 18 (Littre 8.608); Plat. Prot. 322a8; Arist. Hist. An. 535a32; PHerc. 19/698 col. xxvii 6 (Monet). Cf. Cic. ND 2.149: [sc. lingua] "vocem inmoderate profusam fingit et terminat atque sonos vocis distinctos et pressos efficit eum et dentes et alias partes pellit oris." On descriptions given by other Latin authors on the articulation of the voice, see Biville 2001. Articulus in the sense of "joint" is found at DRN 3.697. 28Cf. Hp. Carn. 18 (Littre 8.608): r\v be ut]r\ y^coaandpOpoinpocfiaXXovoa ?Kdoxox?, ctuk dv aatpEGx; 8ia^?yoixo, aXX' r\ ?Kaaxa qvaei xd (xovoepcova the tongue does not ("if articulate by touching [sc. the palate and the teeth] each time, the person does not speak clearly, but utters, as they all are by nature, mere sounds [my italics]," trans. Potter). See also Mar. Vict. GL 6.4.20-21, who opposes vox articulata to vox confusa, the latter being that which sends forth nothing but the simple sound of the voice (quae nihil aliud quam simplicem vocis sonum emittit), and Koenen 1999a, 460, n. 76, on the opposition. Describing the groan of someone suffering intense pain, Lucretius describes the semina vocis being borne from the mouth as glomerata, "all bunched together" (3.497), i.e., not articulated.

CRAFTED SPEECH IN DE RERUM NATURA

541

and speaks directly where Lucretius collapses such a chain of reference of words traveling through the air (4.558,580). Moreover, the tongue is a words as synfashioner of words, implying that we should understand with the articulations that issue from the mouth rather than onymous of the things whose form is copied. At the same time, in the discussion verborum (4.574) so that it appears that echo, Lucretius refers toformae the verbum is itself the thing whose form is reproduced rather than a form shaped in accordance with something else. In this context, Lucretius the word itself is For, although implies yet another level of imitation. it is back into circulation repulsed by the rocks, in being propelled even what is redoubled Yet, on closer inspection, (iterabant, 4.579). pelled is described as words or voces, with the result that there still seems to be an equivalence between the articulations, words, and their forms. A of a word is not at one level of remove from the word understood form as a copied object but is, rather, synonymous with that which it imitates. The formae verborum thrown back faithfully from the rocks are interwith the word, suggesting that, perhaps, the word is nothing changeable but its form, as opposed to an object that is qualitatively different from the simulacrum that flies off of it. Of course, we still do not know what it is a form or a figure of. of the verbum as both created The interpretation object and simulacrum is supported by Lucretius' discussion of the voice's ability to reproduce itself in order to reach the ears of many listeners, although we find here the problem of the forma further complicated: in multas igitur voces vox una repente diffugit, in privas quoniam se dividit auris obsignans formam verbis clarumque sonorem.

(4.565-67)

Thus one voice at once disperses into many voices, since it divides itself up for separate ears, stamping its form and its clear sound onto the words. The splitting of the sound stream occurs in Epicurus' account of hearing (Ep. Hdt. 52-53), but, as Edward Lee has noted (1978, 55, n. 12), it is a than the one under discussion here. Neverwholly different phenomenon of articu? the specificity that it presents elucidate theless, the problems the segmentation of the sound current lated speech. Lee understands described by Epicurus not as a mechanism for delivering the same sound to multiple listeners but rather as the point at which differentiation in speech would within the sound stream is produced: thus segmentation addresses of phonemes Lee's interpretation be the production (1978,31). a real lacuna in the Epicurean account, for at no point there do we find

542 a clue as to how the variation

BROOKE HOLMES

of the voice to produce music or speech, so dear to Lucretius, comes about. Whether or not Lee is right, his transfer of what in Lucretius is the work of the mouth to a process of spontaneous segmentation captures the absence of agency in Epicurus and his on sound as the quality of a body. If we do understand segmen? emphasis tation here as a means of reaching multiple listeners, there is no place at all for difference within the sound stream to be created. Epicurus' talk of In any case, homoiomerous particles supports this latter interpretation. the sumpatheia that the sound particles are said to maintain with one that each conanother and the "distinctive unity" (henotes idiotropos) both pertain to that body qua serves vis-a-vis the sound-producing body referent.29 model for vision and hearing, accord? Epicurus rejects Democritus' are stamped by an object onto to which impressions (apotuposeis) ing pieces of air and, thus molded, travel to the listener or viewer (DK68 in A128). For Epicurus, sound particles, like eidola, are simply expressed will produce the form appropriate to the specific akoustikon pathos they modification for his (Ep. Hdt. 53). While Lucretius adopts the Epicurean which simply peels off the object, the idea of molding re? simulacrum, turns in his account of hearing because the unit of sound is a word. Two ensue from this shift to articulated speech. First is the complications of the vox. Once the object to be figured by the lips and strange agency (the verb obsignans tongue, the vox, as the subject of the participle to "affix a seal to"), usurps their role and itself meaning technically stamps upon words a form and a "clear sound."30 At the same time, as the itself by grammatical subject of diffugit and se dividit, it reproduces replicas as though it were an spontaneously splitting into numerous automatical effluence, not of the sound-producing body but perhaps of the original articulated word. It is clear that it is useless to maintain any distinction between voces and verba, since the articulated vox, once divided, is simply lots of little articulated voces, that is, words. The moment at which the vox multiplies is crucial for the relationship between the simulacrum and the word, for it is at this point when that which is

29For Lee, the sumpatheia preserved reestablishes the continuity of the auditory stream (1978, 32-33), while the "distinctive unity" communicates timbre. 30Cf. 2.581, where knowledge is being pressed upon the memory: "illud in his obsignatum quoque rebus habere / convenit et memori mandatum mente tenere." See also the pseudo-Aristotelian On Things Heard 801b3-6: d8i)vaTov yap juf| xzXivx; totjtcov eivai aacpei*;, Kai oxav uf| 8ir|p6p(0|i?VCL)v cpcovac; xaq KaOdrcep xac,xcov8aKTuA,i(ov ocppayiSac;, 5iaT\)7C(o6(oaiv aKpiPfix;. Epicurus himself uses ?va7ioa(ppay{^o|iaiat Ep. Hdt. 49, but this concerns his rejection of the Democritean theory.

CRAFTED SPEECH IN DE RERUM NATURA fashioned henceforth

543

by the mouth generates its own repetitions. These repetitions Yet a distinction has to carry out the work of communication. the form fashioned be made, namely between by the mouth and its Deand generated which are both re-fashioned automatically. replicas, the equivalence at this level between the word, the voice, and the spite form, we are still faced with two distinct stages, the first being anterior to the process of replication and involving an act of creation. the between the coincidence seem to reinforce Several passages the formae word. First, visual simulacrum and the replicated/replicating verborum were thrown back from a rock in the same way that a mirror in the case of an (4.290-301), although, repulses a visual simulacrum echo, the word is given in its proper order (ex ordine) rather than reversed. Furthermore, we see stray words that, having failed to reach the deears of a listener, bounce around as echoes and, upon becoming formed, create the illusory music of nymphs and satyrs, in the same way that stray simulacra cause us to believe in monsters. Lucretius explicitly albeit in order to at 4.595-614, these replicas to simulacra compares their differences, and he goes into greater detail about the splitting: signal praeterea partis in cunctas dividitur vox, ex aliis aliae quoniam gignuntur, ubi una dissiluit semel in multas exorta, quasi ignis saepe solet scintilla suos se spargere in ignis. ergo replentur loca vocibus abdita retro, omnia quae circum fervunt sonituque cientur.

(4.603-8)

Besides, a voice is divided in all directions, since voices engender other voices when one, having sprung forth, once bursts into many, just as a spark of fire is often seen to scatter itself into fires of its own. And so places hidden from sight fill with voices, and are astir on all sides, and teem with sound. voice dividing into As a result of engaging in this auto-replication?one word is not in turn, allow more to be born (gignuntur)?the many, which, which must be as the visual simulacrum, subject to the same limitations bred from a single source. This distinction explains why it is so difficult to between the articulated voces, the verba, and the formae distinguish are qualiverborum. Once created, they replicate, and these replications tatively no different from the original, like flames from a single fire. The original is no more, no less real than the copies. Yet in re-examining this passage, we may note a certain slippage. Places hidden from sight teem with voices, but these voices are no longer

544 classed

BROOKE HOLMES

as words: they are only sonitus, sound that may be taken for the haunting music of Pan.31 And this returns us to the second complication of articulated speech, namely the imprinting of form onto the word. Like voice is threatened articulated the simulacrum, by transit, which may cause it to lose its identity, however construed. Unlike the iconic simula? a similar, although that produces crum, the result is not a distortion altered image, as with the round tower (4.353-59). Rather, the effect of shift in the nature of the bit of traveling the air causes a qualitative matter itself. No longer a word, it is only a piece of sound signifying the loss of a word. Hearing a word damaged in transit, you sense the sonitus, which is to say it has attracted your attention, but you cannot distinguish A word that goes unheard perits meaning (4.560-61).32 (internoscere) ishes in vain (peritfrustra, 569), unless, perhaps, it is beat back as an echo, which may return either the forma verbi or a sound that only taunts (frustratur) the listener with the image of a word (imagine verbi, 571 ).33 At this point, the interplay between the visual and the auditory becomes between the word's forma, downright maddening: what is the difference 31For the topos evoked here, see Buchheit 1984, 141-47; Gale 1994, 133-36. It is suggestive that in the description given by Lucretius of the afterlife of words echoed in the hills, he imagines that, no longer functioning as signifiers, they turn into music. As I point out above, the deformed word is something qualitatively different. Music becomes the remainder of speech. While Lucretius' plea to Memmius/'ta/ac ne ventis verba profundam" (4.931), may revisit a common theme, we may also imagine it literally, as a plea for the poem to be heard lest it dissolve into mere sound (pure song?). It is fascinating that Lucretius dwells on the failed word in his discussion of distortions of hearing; cf. Sext. Emp. Math. 208, where the type of "hearing falsely" (pseudos akouein) made analogous to the distortion of the round tower concerns a judgment regarding the volume of sound. 32 "ergo fit, sonitum ut possis sentire neque illam / internoscere, verborum sententia quae sit." Words that have been "dashed up" against solid objects (allisa, 4.570) similarly draw attention to themselves but do not deliver any meaning. Cf. 4.613-14 where an [articulated] voice, blunted (obtunditur) and confused (confusa), causes us to believe we hear a sonitus, rather than a verbum. For obtundere and confundere with visual simulacra, see 4.355 and 5.580, respectively. As I note above, they still transmit an approximation of the object's form. 33Cf. 4.221, where he is extrapolating the argument about the simulacrum to the other senses: nec variae cessant voces volitare per auras; see also 6.927: nec varii cessant sonitus manare per avris. Despite the auto-replication of the articulated voice, these must refer to sounds that are continually emitted by objects, since unheard words dissipate if they do not echo back. The use of voces at 4.221 rather than sonitus, then, is haunting. The echo captures the point where the word escapes the confines of the created world established between a speaker and listener; it is the identity of the word once it has gained the autonomy of a sustasis, a fragment of the real no longer bound to it. These unceasing voces hint at intelligibility without meaning, the point where that created according to the fixed laws of one system has been let loose in another, governed largely by contingency.

CRAFTED SPEECH IN DE RERUM NATURA which

545

its identity qua word, and its imago, which is only a preserves sound? What is the latter's unfulfilled promise? teasing a of sound. Hearing them involves Words are not a subspecies different act of perception. A compound form, forged from shaped sound its own simulacra (and of particles, the word is capable of generating These simulacra, like back into its component too, disintegrating, parts). the "original" shaped by the mouth, bear a "clear sound" and a form, a use oiformatura and figura at pair that recalls the apparently pleonastic 4.556. It is unlikely that a clear sound refers to the nature of the sound since this is a constitutive quality that is not, like particles themselves to to distortion. Yet it may, like the formatura, correspond form, subject that distinthe specific "sound of the sounding" . . . sonanti) (sonitu by guished a given word in the letter-atom analogy and was determined the particular order of the elements. In the analogy, of course, the word has a dual identity, both the sound of its sounding and the object denoted (1.826). And, as we have seen, if a word loses its distinct pattern, it loses, too, the specificity of the res, at which point it ceases to refer to anything It collapses but its own inability to make reference. back into sound, but understood not as sensory information conveying, say, hoarseness, the loss of meaning (sententia).341 would venture that, on the one hand, of the shape of the object flickers behind the notion of the something word's figura, a word that is a regular synonym of the visual simulacrum in Book 4.35 The imago verbi, then, is a tautology, in the sense that the 34Cf. 5.1052-55, where a hypothetical nomothetes attempts in a pre-linguistic stage to teach a purely conventional language: "nec ratione docere ulla suadereque surdis, / quid sit opus facto, facilest; neque enim paterentur / nec ratione ulla sibi ferrent amplius auris / vocis inauditos sonitus obtundere frustra."Since the names are conventional and should be established by common consent, they are meaningless when taught by a single person, and those hearing them may as well be deaf. Lucretius acts here as though phonemes are only meaningful in the compound form of the word. Note the playful tmesis at 4.562, inque pedita (in- only becomes meaningful when rejoined Wxthpedita),with Hinds 1987 on 1.452. See Arist. Int. 16a 20-28, b30-34, and cf. Plat. Crat. 426c-27d, and passim. Lucretius' decision to bypass any notion of an independent meaning for sound is consistent with Philodemus' attacks against the kritikoi and Crates.Thus he keeps sensing to the sound and discernment to the level of meaning, where Crates speaks of discerning (diagignoskein) the material differences (phusikai diaphorai) of poetry; see On Poems 5, PHerc. 1425 col. 27 13-21 (Mangoni); I owe this reference to J. I. Porter. On Crates' provocative mixing of the sensual/sensible and the intelligible, see Porter 1989. 35 Figura is often used interchangeably with forma to denote the irreducible shapes of atoms (forma, e.g., 2.334,723; 3.32; 4.27,678; 5.440;figura, e.g. 1.685; 2.335,341,484,682; 3.190, 246; 4.648; 5.440; 6.776), and both words can also describe perceptible form (2.27682). But figura alone is a regular synonym of simulacrum (1.950; 4.27,46,109,158,317,738; cf. 4.104, where forma seems to denote simulacra repulsed by a mirror).

546 word

BROOKE HOLMES

is no longer an image of anything but itself, which is to say a form that fails to translate, even badly, into the visual. The damaged of a word is not much of an image at all. Thus, on the other hand, image to the iconic, of the specificity of verbal form, its irreducibility something distinin phrases like formae verborum. Bailey (1947,3.1248) lingers, too, from figura by arguing that the former refers to the guished formatura while the latter of the voice-cluster/word, atomic arrangement refers to the external shape of that cluster as it is recognized by a listener with the language. In other words, the shape of a material acquainted in a conventional from the form it possesses object is distinguished of symbols.36 Given that words seem to function primarily for system internal Lucretius by making listeners see, it is easy to see how figura and forma might straddle the auditory and the visual, hovering at the moment the and the word nearly collapses the visual simulacrum analogy between into identity, smoothing the gap between the form of a word and what it of these two spheres is stressed by the shows. The interconnectedness What of servare and the leonine rhyme formaturam/figuram. repetition exactly, then, is it to perceive II. WORDS For Epicurus, the condition a word? AND THEIR IMAGES

the senses are irrational (alogoi, D.L. 10.31)?this being Lucretius embraces the doctrine of their infallibility?and Thus, the eyes see based on the simulacra they receive, wholeheartedly. is made about what the subsequent but no judgment imago represents for "the eyes cannot know the nature of without the mind's intervention; oculi naturam noscere rerum, 4.385). If this interven? things" (necpossunt tion is necessary for knowledge, it also carries risks: with a strange turn of phrase, Lucretius, in the very next line, refers to the capacity of the mind to know the nature of things as its fault (animi vitium hoc, 386), thereby marking the act of judgment or inference as an act poised to fail. Epicurus himself describes the addition of doxa as a mental movement (kinesis) whose trajectory thus discloses the space in which both error and knowl? become possible (Ep. Hdt. 51; see also Sext. Emp. Math. 7.210). external Conversely, given that a sensation is set in motion by something to the body, there is no room for it to add or subtract anything at will,37 edge 36One may speculate that Epicurus or Lucretius would imagine something like an auditory prolepsis of the word, i.e., a generic phonetic form allowing a listener to recognize individual variations; PHerc. 19/698 col. xxvi-xxvii (Monet) may support such an idea. 37oi)iE yap txp' atrcfjs; oike txp' eiEpoi) KivnGeiaa 5t)vaTai xi jcpoaBeivai r\ dcpe^dv (D.L. 10.31).

CRAFTED SPEECH IN DE RERUM NATURA

547

with the first a formulation that, as we will see, has much in common of of language. Lucretius inherits from Epicurus this vocabulary stage in dreams in terms of addition and excess, accounting for our deception the opinions which we ourselves add (opinatus animi quos addimus ipsi, what is clear, movement introduces, alongside 465). This self-generated what is also doubtful apertas / ("nam nil aegrius est quam res secernere ab dubiis, animus quas ab se protinus addit," 467-68). is the risk of erring, we know that the cost of knowledge Although Lucretius does not specify how the mind acts on sense data.38 While there in us from sensation to recogni? can be no seeing without the movement the best hope of making true to knowledge, and from recognition tion,

the of perception, claims about the world requires optimal conditions "near view." Thus, towers that only appear round as a result of damaged simulacra are contrasted to things that are round "from up close" and The adverbs coram and vere enjoy a symbiotic relation? "truly" (4.362). objects ship: only the "near view" favors claims about macrophysical otherwise must be held in abeyance (steremnia). Any thing apprehended to transarises to confirm or refute it.39 Returning until the opportunity unlike the deformed that a damaged word, mitted words, I suggested dead-ends simulacrum, qua word. But what of the word properly heard? David Glidden has written "there is no such thing as Epicurean This is indeed uncertain terrain, and there is little semantics" (1983,204). consensus principles among scholars on even the most basic Epicurean of language.40 Any casting about for a coherent theory of the relationship lacuna in the between words and things must confront a troublesome in our evolves. For we have solid information of how language story sources on the origins of words and frequent reference to linguistic error via language but very little evidence of how and false beliefs transmitted has been the Central to the discussion nomination becomes fabrication. 38In discussions of perception and perceptual error, commentators typically make reference to the idea of prolepsis, which I discuss above in the context of language; see Bailey 1928, 244-48; Long, 1971, 118. Lucretius primarily exploits the dichotomy senses/ mind and relies, as I emphasize above, on the moment when the mind "adds"something. For Epicurus, this is a moment or space of difference, dialepsis (Ep. Hdt. 51), a "differential turn," to tropon dialeptikon (fr. 34.22 Arrighetti). This "other" movement at Ep. Hdt. 51 must correspond to the act of judgment. See Long 1971,118; Furley 1971; Asmis 1984,146-66. 39Sent. 24; Ep. Pyth. 85-88. Correct reasoning under optimal circumstances fails under the heading of epilogismos, on which see Sedley 1973, 29-34. What cannot be confirmed permits the coexistence of multiple explanations; see, e.g., 5.509-32. 40In this discussion, I have drawn on Schrijvers 1970, 91-128; Long 1971; Sedley Goldschmidt 1978; Glidden 1983; 1985; Asmis 1984, 19-80; Long and Sedley 1987 1973; 1.87-88; Everson 1994a; Barnes 1996; Hammerstaedt 1996; Porter 1996.

548 controversial

BROOKE HOLMES

notion oiprolepsis, which serves as the primary mechanism Laertius words to reality. That prolepseis are classed by Diogenes relating and feelings (pathe) as one of the criteria of with sensations (aistheseis) less truth in the Canonic (10.31) suggests that they should be understood as beliefs, even correct ones, and more as traces of our experience of the which we draw upon to make and evaluate claims about it.41 That world, to protect are often caught using prolepseis strategically said, Epicureans it claims from challenge, even when these extend, illegitimately key would seem, beyond the evidence available to direct perception.42 The assimilation of the word to the icon that we have been tracing in Epicurean semantics in Lucretius with the preoccupation resonates of what exists, is treated as the designation with nomination. Language that is, regularly recurring atomic composites.43 This promise of speech as an act of deixis has fueled a critical indeterminacy about where the that primarily underlie words" reside. These underlying things "things as "general concepts," are often understood to be prolepseis, imagined and called to mind when one built up from reliable sense perceptions Elizabeth Asmis, has been resistant hears a word.44 Glidden, along with

41See D.L. 10.33: Kai xo 8o^aoxov anb 7ipoxepo\)xivoq evapyovc;fjpxnxai . . . . Cf. Everson 1994a, 102-8; Barnes 1996, 209-20. 42On the strategic use of the prolepsis of a "good poem," for example, see Porter 1996, 626-28. See also Cic. Fin. 1.31 on Epicureans who try to shore up the claim that pleasure is to be sought and pain to be avoided by making it an idea (notio) that is natural and that has been implanted in our mind (naturalem atque insitam in animis nostris; cf. Cic. Nat. D. 1.17, insitas eorum vel potius innatas cognitiones habemus) and thus immune to argument. The prolepsis of a god presents similar problems. In Lucretius, for example, the two qualities that are indispensable to the Epicurean idea of a god, namely blessedness and immortality (see Ep. Men. 123-24), are (correct) inferences (5.1175-82; cf. Cic. Nat. D. 1.17); see Asmis 1984, 74-79. 43 Long 1971,127; Glidden 1983,203-9; Asmis 1984,25; Porter 1996,621. Sextus tells us that Epicurus had no use for grammatike (Math. 1.49,272). The fragmentary remains of Book 28 of Epicurus' On Nature do show him engaging with questions of the relation of complex language to reality; see Sedley 1973. 44The definition in Diogenes Laertius, far from securing the kind of clarity that the concept itself should guarantee, behaves more like a potentially interminable chain of signifiers:prolepsis is "a direct apprehension (katalepsis), or a correct opinion (doxa orthe), or a conception (ennoia), or a universal 'stored notion' (katholike noesis), i.e. a memory (mneme) of that which has appeared frequently externally" (10.33). The problem is made worse by the predominance of Stoic terms. Shortly thereafter, he seems to call prolepsis the "thing which first underlies the word" (to protos hypotetagmenon). Reference to what first underlies a word occurs in a famous caveat about how to make judgments and inferences at the beginning of the Letter to Herodotus (37-38 = 17C Long and Sedley: rcpcoxov ow xa (lev cb av i)7coxexayueva xoi<; q>96yyoi<;,fHp65oxe,8ei ei^ncpevai,O7tco<; xa 5o^a^6u?va r\ ^nxouueva

CRAFTED SPEECH IN DE RERUM NATURA

549

to understanding this underlying thing as a mental entity, preferring to in the world, whether it be see the word rather as designating something a solid object or the mere eidolon of what lacks any solid instantiation, While what primarily concerns such as a god or a Centaur (1983,198).45 him and Asmis is what a word means rather than what happens when with the one hears a word, the idea that a word provokes an engagement external world accords with what little evidence we have from Lucretius. facilitates the creation of such An emphasis on simple nomination concerns the recogin fact, prolepsis frequently a relationship, although, abstract entities, nition or imagination not of objects but of seemingly own account of naming, we in Diogenes' such as justice.46 Nevertheless, are dealing with macrophysical morphological objects with discrete,

Kai r\a7topo\)U?voc eiq excofiev xauxa avayayovxeqe7iiKpiveiv, ur|aKpixarcavxaf|uiv eiq arceipov d7ro8?iKv/6o\)oiv Kevotx;cpGoyyoix; evvonjia Ka9' eraaxov r\ avdyKn yap xo rcpcoxov e'xo))iev. cpGoyyov P^e7i?a0ai Kai ur)9ev drcoSei^eax;Jipoo8?ia6ai, eutep e^ojxev x6 ?nxot)|j,?vovr\ Kai dc7ropoiL)uevov 8o?a?6u?vov ecp'o dvd^o|iev, "First, then, Herodotus, we must grasp the things which underlie words, so that we may have them as a reference point against which to judge matters of opinion, inquiry and puzzlement, and not have everything undiscriminated for ourselves as we attempt infinite chains of proofs, or have words which are empty. For the primary concept corresponding to each word must be seen and need no additional proof, if we are going to have a reference point for matters of inquiry, puzzlement and opinion" [trans. Long and Sedley]). Sedley (1973,14-17) has argued that proton ennoema should be understood as a precursor of prolepsis, which he argues was a later addition to Epicurus' technical vocabulary.This passage, then, has lent support to interpretations that take prolepsis to mean the idea?ennoema?that subtends a word; see Long 1971; Sedley 1973; Everson 1994a. Note, however, that in the commentary on Ep. Hdt. 37-38 in Long and Sedley 1987, the absence of the term "merely reflect[s] Epicurus' concern in the opening moves of his physical exposition to appeal to the most general possible considerations, leaving the more heavily theory-laden terms to emerge in due course" (1.89). For speculation on the meaning oi proton ennoema, see Asmis 1984, 31-32, with bibliography (n. 31). 45See also De Lacy 1939, 85; Asmis 1984, 26-30; Glidden 1985. Glidden and Asmis attempt to make sense of evidence from Sextus and Plutarch (Sext. Emp. Math. 8.13, 258; Plut. Adv. Col. 1119f-20a), who both claim that Epicurus did not allow anything like the Stoic lekton to mediate between what is real and the sign but dealt only with the thing and the utterance. On the Stoic lekton, see Frede 1994. For attacks on Glidden's position, see Everson 1994a; Hammerstaedt 1996. Barnes has also argued that a word should be under? stood as a thing rather than a concept (1996, 219), although he differs from Glidden and Asmis in that he sees prolepseis as true beliefs that make the use of words possible. 46On "the just," see Epic. Sent. 37, 38. Lucretius' use of notitialnotities is consistent with this tendency. He speaks, for example, of a familiarity with the true (4.476). These qualities and things are still objective, even if, like the gods, they are only accessible to direct mental perception. Things which are known analogically or inferentially, such as atoms, cannot have their own prolepseis, for they do not produce any kind of sensible record; see Sedley 1973, 21.

550 identities,47 a schema model of recognition uttered, immediately oi prolepsis, since the each name is clear"

BROOKE HOLMES itself well to Lucretius' primarily visual and knowledge: "as soon as the word 'man' is its delineation (tupos) also comes to mind by means senses that lends

observes Diogenes a cow by means of prolepsis and that we could name nothing if we had not already learned the tupos, again by means of prolepsis.48 While the sense of tupos is no less evident here than that of prolepsis?is it a in the mind, or an eidolonl?what is key here is that this barepattern bones model of nomination sets up prolepsis as a means of engaging with in their absence via some kind of accessible form of them.49 objects 47He is following Epicurus here in his morphological emphasis; see Sext. Emp. OP see 2.25, where man is defined as xo xoiouxovi uxSpcpcouxx ?u\|/i>xia<;; also Math. 7.267. (iexd Cited in Asmis 1984,45-46. 48 Prolepsis here seems to be the faculty that is instrumental in summoning a form of the object as well as the form itself. See Asmis 1984,63-80. On the use of the -sis ending to indicate faculties, see Sedley 1973,33, on epilogisis. Compared, then, with a -ma noun such as ennoema, prolepsis lays more stress on this faculty. It thus registers less satisfactorily as an inert "thing" in our possession, although it may also denote what is accomplished or gained, as aisthesis can mean both the faculty of perception and the perception itself. Bailey recognizes this but wants to limit the meaning to the "compound image" that serves as the basis of an "act of anticipation" (1928,562). The real difficulty is knowing what the role of the tupos is. Epicurus himself speaks of xt^rcoi as 7tpo?i^[r|](p6xe<;an apparent periphrasis for prolepseis (34.28 Arrighetti = 20C Long and Sedley, whose text I follow), and tupos is a synonym for eidolon (Ep. Hdt. 46; cf. 36; 49); cf. Phld. On Poems 5, PHerc 1425 col. 30 2933 (Mangoni): dv 8e 8id xo-ulxov\x6v(oq oi(bu?0a xaq I K[po]Xr\\\f?iq ektvuovoQcli, I Ttdvxa On 7c[ap]a0exeovxcoiI yevei, [aXX'<ju]xoiq dpiGuoic;. tupoi, see Goldschmidt 1978, 156-64. On the relationship between memory and tracing or imprinting, see also Diog. Oion. fr. 9 col. iii 6-col iv 2 (Smith); DRN4.428-31; 6.995-97; Plut. Mor. 735a (= DK68 A77 = Us. 326). Not all these cases are about prolepsis, but they do see our experience of the world as carving out an increasingly subtle receptivity to it. 49A second concept from Epicurean epistemology becomes relevant, then, namely [phantastike] epibole tes dianoias, which permits the mind to seize upon simulacra at will. Epibole generally implies a form of concentration, e.g., listening rather than simply hearing. The image grasped is valid as a criterion of truth (Sent. 24; D. L. 10.31). Bailey (1928,42831; App. III) made epibole tes dianoias the means by which atomic reality was perceived, as though it were sensible rather than accessible only to reason; see also Thury 1987,282-83, who goes so far as to argue, wrongly I believe, that the poem itself can provide a prolepsis of atomic theory. Cf. Furley 1971; Sedley 1973,23-25.1 would agree with Furley and Sedley that epibole tes dianoias only pertains to the mind's attention to perceptible things, objects or eidola of the gods, although I do not agree with Furley that epibole tes dianoias only deals with illusory objects. It deals, rather, with objects in their absence; see Asmis 1984,8691; 124-26. The precise relationship between (phantastike) epibole tes dianoias and prolepsis is unclear; Clement does describe prolepsis as an epibole uepi ti enarges" (Strom. 2.4 = Us.

give the lead: thus what primarily underlies Further down, (10.33, trans. Long and Sedley). that we have learned the form (morphe) of a horse or

CRAFTED SPEECH IN DE RERUM NATURA Notitia,

551

the word that Cicero tells us?albeit in a Stoic context (Ac. used to translate the Greek terms prolepsis and ennoia, and 2.30)?was its collateral form notities appear in De rerum natura eight times. Yet these terms do not appear in accounts of mental processes or perception, nor do they surface in relation to our knowledge of the gods.50 What dominates such discussions, rather, is the simulacrum.Thus, when Lucretius counsels Memmius to develop beliefs about the gods, his appropriate stated goal is to give him the courage simulacra?those envoys of divine form from the gods' bodies to the minds of More importantly, thought in Lucretius his consistent with tranquil mind the (divinae nuntia formae)?borne humans (6.76-78; cf. 5.1169-71). is overwhelmingly visual, both in to seize

as insight and in his explicit, albeit figuration of philosophy limited account of mental processes in Book 4. This is unsurprising, given that the extremely like the correspondfine, mobile mental simulacrum, eidolon in Epicurus on the visual one.51 Hdt. 49), is modeled ing (Ep. However we are to understand the notitia, say, of usefulness (5.1047), Lucretius paints a vivid picture of a world teeming with exceptionally delicate simulacra in his account of mental processes in the latter half of Book 4. Thus simulacra are far in excess of the number of solid, macrophysical objects in our immediate vicinity, and they are readily available to our minds at any time. Interaction with this external, simulacral dimension accounts for our ability to think of things at will, as well as for our dreams, which are fueled by a continual influx of images sourced from this bank. At 4.802-15, of the mind both Lucretius' description

255), and Sedley believed that the notion of epibole tes dianoias was subsumed under the general heading of prolepsis (1973,16). Both are implicated in faculties that operate with the flimsier eidola that penetrate to the mind and are crucial to activities such as dreaming, remembering, thinking, and speaking. Glidden 1985, 187-201, argues that a prolepsis is a specific type of epibole tes dianoias that allows one to perceive general qualities in objects; cf. Hammerstaedt 1996, 234-37. What appears distinctive about [phantastike] epibole tes dianoias is that it is provoked by an act of will: the mind decides (or is told) to pay attention to this instead of that. Prolepsis, on the other hand, is rooted in the impact that things make on us. 50For notitia, see 2.124, 745; 4.476, 854; 5.124. For notities, 4.479; 5.182, 1047. On prolepseis of the gods, see Ep. Men. 124; cf. Cic. ND 1.43, where prolepsis is translated as notio, anticipatio, and antecepta animo rei quadam informatio. 51 Sedley 1973,23-34, gives a useful overview of Epicurean thought-processes, ranging from those that operate kata ton epibletikon tropon or phantastikos, which rely on the presence of images, to those that operate perileptos or theoretikos, which deal with problems that require some distance to be taken from the image. As I note above, Lucretius focuses in Book 4 only on the first two types. On the visual nature of thought in Epicureanism and the role of mental eidola, see also Asmis 1984,105-40.

552

BROOKE HOLMES

itself to receive and preparing (parare) straining to see (contendere) of its active the intertwining the ambiguity of this interaction, captures and absence and the play between and passive dimensions presence is always present through its simulacral avatars; absence is (everything the result of our inability to think the world all at once). The saturation with simulacra also explains what happens when we of our environment hear a word:52 an magis illud erit verum? quia tempore in uno, eum sentimus,53 id est, eum vox emittitur una, tempora multa latent, ratio quae comperit esse, propterea fit uti quovis in tempore quaeque praesto sint simulacra locis in quisque parata: tanta est mobilitas et rerum copia tanta.

(4.794-99)

Or rather is this true? Because in one moment, when we perceive, that is, when a single word is uttered, many moments lie hidden, which reason discloses, so that it happens that at any time and in every place these simulacra are at hand, readied. So great is their speed; such a wealth of them exists. What primarily interests Lucretius here is the operation of two orders of to our perception?this time, the one in which a word calls a simulacrum the one in which the appropriate simulacrum instantaneous?and appears 52See Schrijvers 1970, 99-128; Asmis 1984,120. It is hard to see how reading vox as simply sound here suits the context, since what is at stake is not merely atomic time but also the temporality of simulacra. The spoken word's relationship to the simulacrum is well suited to proving the nature of that temporality in that it takes place in a discrete moment, for articulation introduces difference not only into space but also into time. There is no correspondingly clean way to mark a single moment of seeing. Moreover, as the discussion of hearing in Book 4 made clear, Lucretius is not interested in the transit of the sound particle: the splintering of the sound stream discussed by Epicurus is not a part of his exposition. Finally, the visual simulacrum is the model for the extraordinary speed of all simulacra, for it is fastest, as Lucretius recognizes at 6.165-66,183-84. On the use of vox for verbum, see 4.562, 565, 568, 577; 5.337. Schiesaro 1994 also understands the word (for him primarily written) as giving rise to a visual simulacrum, although he is not consistent in his presentation of the process. On the one hand, he argues that the word accomplishes this by giving the mind instructions to create an image from free-floating atoms (87-88), a possibility denied by Lucretius in the lines cited above (natura does not make simulacra ad hoc); on the other hand, he follows Schrijversin making the mind attend to the appropriatesimulacrum (88-89). Thury 1987 assumes interaction between language and vision but does not detail it. 53The MSS give the nonsensical consentimus. Munro, Merrill, Ernout, Martin, Biichner, and Schrijvers (1970, 102) print eum; Bailey and Giussani, quod. See Schrijvers 1970,102-4.

CRAFTED SPEECH IN DE RERUM NATURA

553

is summoned from the bank.54 What we may infer is that Lucretius that words (strictly speaking, names) call forth simulacra, which, imagines come to the mind like the tupoi in Diogenes, instantaneously functioning of the listener. Further support for this is given at 4.785, where Lucretius, that allows us to see whatever we having not yet specified the mechanism them all [sc. asks ironically, "does nature make and prepare wish, at a word (sub verbo)V'55 While he rejects the idea that a simulacra] its appear? from scratch, in associating simulacrum would be generated 4.794. Thus, the word is a provocation ance with the word, he anticipates and hearing words has strong to attend to an object (via its simulacrum), of verbal and visual affinities with seeing that object. This intertwining of figura and forma in the form sheds light on the overdetermination account of hearing. To return to the letter-atom analogy, even a minor in the elementa, say from mare to terras, summons the simulacrum change of a different res. In this stripped-down, deictic model, one fashions specific words because one wishes to bring certain simulacra to the listener's attention. But to shift attention to the speaker, why do we utter the words that we do, particularly in cases where the object is not in front of us or does not exist (qua object) at all? The problem of why we attend to the images we do is one that dogs Lucretius in this latter section of Book 4. It involves the introduction of words such as libido (4.779) and voluntas (481), words that reinsert the mind between things and words and hint at a more serious rupture in the relationship between the world and lan? with ratio, voluntas or libido is something added to the guage. Together world, agitating in the space of choice that allows us both to judge and to misjudge, to act and to fail.56 Yet even if we strip away this act of choice, and the subsequent act of creation, words are not effluences: the differ? ence between their form and the object represented cannot, in the end,

54On "atomic time," see Ep. Hdt. 62 and Long and Sedley 1987,11.45-46. 55Most modern editors have chosen to read sub verbo as equivalent to sub iussu, although see Giussani (1896-98) ad 4.783. Schrijvers 1970, 95-98, convincingly rejects the interpretation sub iussu as unfounded. See also Godwin (1986) ad 4.549-84. 56It thus produces what Michel Serres calls "a small, local diagonal escaping from the monotone and from the saturated whole" (2000,146). Lucretius' discussion of free will is, as is often recognized, deeply problematic; see his explanation of the swerve at 2.251-93, and especially 271-83 on voluntary vs. involuntary motion, and cf. the explanation of motion at 4.877-91, where simulacra of motion (simulacra meandi) strike the mind and give rise to voluntas. See also Epic. fr. 34.21-22 (Arrighetti) = 20B Long and Sedley; 34.2630 (Arrighetti) = 20C Long and Sedley, and the discussion in Long and Sedley 1987,1.10712. On the influence of desire and habit on perception, see also 4.962-86.

554 be ignored.

BROOKE HOLMES

Lucretius' account of the origins of It is worth re-examining in order to understand defines the work of how this difference language of language. and opens the door to the autonomy naming III. THE NAMING ANIMAL

In Epicurus' first stage of language, humans make sounds to which noth? or belief?has been added. The world is proing?thought, opinion, of to translate the impression cessed by a body functioning automatically the sensible to register/reproduce objects into various sounds, that is, world in the medium of the voice: o0ev Kai xa ovojiaxa ?^ apx^c, un 0eaet yevea0ai, aXX' amac, xaq yvoeic, xcov Ka0' EKaoxa ?0vr| i'8ia nacxoxxjac; naQr\ Kai t5ia Xaupavouaaq dvBpcoTcwv (pavxda|iaxa iStax; xov depa ekjieutieiv ax?M,6u?vov txp' ?Kaaxcov xcov 7ia0cov Kai xcov(pavxaojidxcov, ax; dv rcox?Kai r\ napa xoxx;xonovq xcov?0vcov 5tacpopd (Ep. Hdt. 75 = 19A Long and Sedley) f] [Us. ?i7i MSS]. And thus names did not first come into being because of convention, but human natures themselves, each according to its individual race, suffering particular affections and receiving particular images, sent forth air formed in a particular way by each of these affections and images, so that differ? ence arose according to the locations of the tribes. that This automatism, to the extent that it excludes choice, guarantees in the relationship between objects and the sounds that they provoke Lucretius maintains this different races is natural and non-arbitrary.57 strong link between natura and the exercise of necessity:58 At varios linguae sonitus natura subegit mittere et utilitas expressit nomina rerum . . .

(5.1028-29)

57See also Ep. Hdt. 75, for Epicurus' formulation of the first stage of compulsory human development understood more generally (aXXh ur|v i)7ioX,r|7ixEov xfjv cpuaiv Kai 7toM,a Kai rcavxoia imo auxcov xcovTrpayudxcov 8i8ax6fjva{ xe Kai avayKaaGfjvai). In the second stage, reasoning refines these advances and adds new discoveries. While the latter two stages of development in the account of language in Ep. Hdt. 76 clearly involve convention, there has been much debate about how the relationship to phusis is maintained: see De Lacy 1939; Vlastos 1946; Konstan 1973, 46-48; Sedley 1973, 18, n. 91; Schrijvers 1974; Pigeaud 1983, 127-29, where he examines the ambiguous status of the ethnos, which he calls a donne sociobiologique that guarantees the existence of a "thesis naturelle" within each tribe. 58See Brunschwig 1977, 160-61; Campbell 2003, 294-96, for an overview of critical interpretations of natura (Nature or human nature).

CRAFTED SPEECH IN DE RERUM NATURA But nature forced them to send forth various sounds of the tongue And usefulness expressed the names of things.

555

Where "we" were the subjects of the verb emittere at 4.550, at 5.1029, "we" are only the implied subjects of the verb mittere, compelled to send of ourselves out into the world. Speech is distinguished, as we something have just seen, not by its ability to mirror but by its work of translation. To the extent that Lucretius preserves the stimulus-response model that is crystallized air emitted is shaped directly by Epicurus' grammar?the and the affections themselves ucp' by the phantasmata (gteXX6\ievov eiccxaxcov xcov 7tcc0cov Kai xcov (pavxaajidxcov)?there is a space for these varii linguae sonitus to refer beyond the body that produced them in order to represent the original stimulus. By denying that names (onomata) are generated by thesis, Epicurus makes it clear that these first sounds mark objects precisely because objects provoke in a repetitive manner the same kinds of reactions in a given ethnic community in a given In its earliest stages, language environment.59 is already talking about and this something exceeds the sound-producing something, body. The difficulty of involving primitive speech in acts that go beyond self-reference is more apparent in the Lucretian formulation, where the varii linguae sonitus, at first glance, seem identical to the cries emitted by animals in response to different sensations Indeed, the post(5.1087-88). of the nomina rerum to the second stage of language makes it ponement clear that Lucretius envisions two distinct phases. With the intervention of utilitas in the second stage, sounds come to indicate objects, rather than simply expressing some state of the body, just as the work of the daedala lingua results in sound referring to the external world. Both utilitas and the daedala lingua establish a relationship of utilitas at the dawn of identity that is not secured by iconic resemblance, language, the daedala lingua in the quotidian act of speaking.The problem is that this relationship can only be established if early speakers observe a pattern of repetition. The object comes to be designated by a sound because it consistently that sound. While this is not arbitrary, produces 59In the account given by Proclus, too, this first stage is explicitly the assigning of names: 6 ydp 'ErciKcyupcx; cruxoi xd Oejievoi]e'Oevxo zXeyev,oxi o\>xie7uaxr||j.6vco<; [sc. oi rcpcoxoi Kai oi kivo\)U?voi, cb<; priaaovxec;Kai nxaipovxeq Kai |ruKcb|j,?voi ovojiaxa, aXka (puaiKax; Kai uXaKcouvxeq axevd^ovxeq (In Plat. Crat. 17.5-17 = fr. 335 Us); see also Demetr. Lac. PHerc. 1012 col. lxvii 9-12 (Puglia); Origenes Cels. 1.24 (= fr. 334 Us). As Mackey points out, the reference to existing deloseis in the second stage in Epicurus means that the first stage must have produced some kind of names (2003, 9). Cf. Konstan 1973, 45-46, whose strained interpretation relies on reading urjwith yevea6ai rather than with Oeoei.

556

BROOKE HOLMES

there is nothing in the effect itself that renders the cause transparent. between touch and sight, When Lucretius establishes the relationship that each sense is moved by the the shared access to shape guarantees same cause (necessest / consimili causa tactum visumque moveri, 4.23233), just as the use of vision as a model for thought renders what is seen But what of the object is most amenable to analogical imagination.60 retained in the sound that it causes a body to produce? The form of the sound emitted should bear witness to a quality of the object that has level: when Lucretius at the macrophysical or that is, its roughness its particularity, the different shapes of the primordia from is guaranteed smoothness, by out once the analogy be? which it is created. This all gets straightened tween words and things at 1.820-29 is in place: the order and position of nothing to do with morphology discusses sound qua sound,

60Cf. Asmis 1984, 106. Sight and touch do technically have their own objects of perception (4.256-68; cf. PHerc. 19/698 col. xx-xxi [Monet]), but it is touch in Lucretius that assures the relationship of the simulacrum to the object. That vision and touch in an analogical sense share shape as an object is laid out in PHerc. 19/698 col. xxv (Monet); see Sedley 1989. That same treatise claims, rather vaguely, that all the senses can be analogically related by form (schema) and does argue that articulated voice has a form analogous to color (col. xxvi-xxvii; see above n. 26), but it is precisely the requirement that voice be articulated in this comparison that is tricky.For what is the nature of the form, which serves as the object of the hearing, when it has been imposed by the speaking subject? The idea that unarticulated sound (either the sound produced by the observer or by the object itself) preserves the form of the object that produces it is difficult: see, e.g., Bailey's translation of (pOoyyoDc; dpKeiv yap xoix; cp-ociKoix; x^pzlv Kaxa xohq xcovTrpayumcov [D.L. 10.31] as "it is sufficient for physicists to be guided by what things say of themselves" (1928,161); Asmis understandably responds "the view that things issue utterances seems to me implausible" (1984,27, n. 21). Thus, one easily lapses into the language of vision, which, as we have seen, does not quite work. Vlastos, for example, uses parallels from Epicurus' theory of vision to explain Ep. Hdt. 75-76: "feelings and impressions directly 'form into shape the vocal sound'; much as in the theory of knowledge, the incoming stimulus can so mold the sensorium that the sense-image will reproduce 'the very form of the physical object'" (1946,52). In his note (52, n. 12), what guarantees similarity is the enkataleimma or residue of the eidolon left in the eye (Ep. Hdt. 50). But, again, this occurs in the explanation of vision where Epicurus says that this residue conveys the form (morphe) of the object. And shortly thereafter (54, n. 16), phantasmata are explicitly said to be "images of objects." While in theory, the phantasma, to the extent that we can ally it with phantasia, covers all forms of sense data (see Sext. Emp. Math. 7.203 for phantasia with all forms of sense perception), the very term betrays a visual bias, as Vlastos' interpretation shows. See also Ep. Hdt. 51; Ep. Pyth. 102 where phantasma is clearly visual and Ep. Hdt. 49 (the stream oieidola produces a.phantasia). On the resemblance between a sound particle (which only preserves a sumpatheia with the other sound particles in the auditory stream) and its source, see Lee 1978. But, of course, this resemblance has nothing to do with the sound produced in response to being affected by the object but rather with the relationship between auditory particles and their source.

CRAFTED SPEECH IN DE RERUM NATURA

557

the letters translates (in a very loose sense that requires the intervention of convention) the order and position of the atoms, which gives things level. But in the beginning, their specific form at the macrophysical those objects systematically words are related to objects because (at least in a given natural environment) produce the same vocal reaction. secured by the daedala lingua rather than This regularity is eventually the object and, indeed, without the need for there ever having been an object. to seeing these early Yet Lucretius, like Epicurus, is still committed sounds as naturally deictic.61 For the figure of utilitas occludes the collec? that relates the sound tive mental action of the early grunters/speakers sutures them back to the object. It appears as the force that impersonally of the While in his refutation resemblance. and guarantees together would have Lucretius argues that the name-giver nomothetes argument needed a notities of the utility of names, in his own story he avoids the circularity one have of the argument (if no one was using language, how could any choice altogether an idea of its usefulness?) by eliminating What Lucretius aims to describe is a paradoxical (utilitas expressit.. .).62 but rather state where humans do not produce sound under compulsion of the sounds were as though the referentiality name under compulsion, The biological of the simulacrum. as the referentiality as transparent of a scenario in which humans confirm this in their depiction analogies not simply to make sounds but to re-present are creatures compelled objects in the world outside of them. They do so not by giving forth copies but by creating variation in the medium of the voice. of birds, dogs and When we are invited to examine the capabilities cattle to emit varii linguae sonitus, the question is not to what use each sound but what it is that forces creature puts its ability to manipulate them (cogunt, 1087) to send forth their different noises. The answer is emotions such as fear, pain, and joy (1059-61). Humans, too, are affected 61The use of nomina rerum at 5.72 suggests that the earliest stage in Lucretius, as in Epicurus, involves names rather than expressive, animal-like sounds; see Offermann 1972, 154; Sedley 1973, 18, n. 91; Brunschwig 1977, 172-74; Pigeaud 1983, 124-25; Dalzell 1987, 26-27; Wigodsky 1995,62, n. 24; Mackey 2003,8, n. 15. Cf. Bailey 1947,3.1486-91 (ad 102890); De Lacy 1939; Cole 1967,61 with n. 3; Schrijvers 1974,340; Snyder 1980,19-22; Glidden 1983,200. For an overview of the debate, see Campbell (2003) ad 1028-29. Also controversial is whether 5.1029 (utilitas expressit nomina rerum) corresponds to Epicurus' second stage. Offermann argues it only reworks the first stage (1972,155); cf. Vlastos 1946,54, n. 17; Schrijvers 1974, 340-46. Much of the confusion stems from Lucretius' attempt to have it both ways, that is, to make names iconic, as I argue above. 62See 4.823-57 for the anti-teleological argument, with Campbell (2003) ad 1046-49.

558

BROOKE HOLMES

from that of the sensus by varii sensus, but this diversity is distinguished that drive animals in that it exists in direct relation to the prodigious diversity of things in the world.63 Of course, as the letter-atom analogy reminds us, this diversity is not absolute: there is not one sound for every arises from the world, difference object. Rather, as in the perceptible of a handful of basic elements that are, first and foremost, manipulation bits of phone. The physiological quirk that permits this manipulation in sound the multiplicity of the simulacra impacting the body. preserves Only humans have tongues agile enough to reproduce (and then recrethe diversity of the natural world through an exploitation of phoate) netic difference, both as it pertains to individual sounds and to their with movejointing together.64 Yet the animal analogy's preoccupation ments that originate outside the body?in Lucretius' specific examples, Molossian hounds and young stallions are driven to express anger or lust?obscures the hints of creative mimesis inherent in the suggestion that the plenitude in the perceptible of difference world might be remade in voice. It invites us to imagine that humans are driven to remake the distinct forms of objects (dissimilis alia atque alia res voce notare, cf. 1043-45) out of what the author of the Hippocratic treatise 5.1090; Fleshes called the "monophonic" 18 = Littre 8.608). If we back(Carn. track to the first analogy between humans and animals, we get an even stronger sense that the impetus to single out objects is somehow uniquely human. In the first analogy (1030-40), animals are introduced to show that each creature exploits its body in order to act in a way that is useful. Even before the tongue is sufficiently in the human child, she developed is driven to point to the things in front of her (praesentia) with her finger, the same way a calf attempts to butt with its head before its horns are fully formed.65 Where again it is an emotion, i.e., anger, that impels the

63 "postremo quid in hac mirabile tantoperest re, / si genus humanum, cui vox et lingua vigeret, / pro vario sensu varia res voce notaret?" (5.1056-58). 64See 5.71-72, "quove modo genus humanum variante loquela / coeperit inter se vesci per nomina rerum,"where variante loquela (from loquor) stresses that the difference within language that allows it to correspond to the world of nomina rerum is phonetic. See Offermann 1972,152; Brunschwig 1977,164; Pigeaud 1983,131-33. 65"non alia longe ratione atque ipsa videtur / protrahere ad gestum pueros infantia linguae, / cum facit ut digito quae sint praesentia monstrent. / sentit enim vis quisque suas quoad possit abuti" (5.1030-33). Campbell (2003) ad 1033-40, notes the difficulties that this analogy poses to Lucretius' evolutionary theories. He had claimed at 4.823-57 that "what is born creates its own use" (quod natumst id procreat usum, 835), so that the question arises as to why children want to point to the world before they can speak if speaking is the result

CRAFTED SPEECH IN DE RERUM NATURA

559

calf, the child is driven to point out objects as best she can because she is aware of a power that Lucretius seems to identify as correspondingly one. Not the power to indicate. The analogy is a complicated human, i.e., is Lucretius likening a child's use of gesture to a young animal's use only to the of its claws or wings, he also makes these gestures analogous of the natural world, sounds that humans produce in their re-presentation of the tongue's natural sounds made possible by humans' exploitation is that the gesturing child, as soon Of course, the implication capabilities. as she overcomes the infantia linguae, will not point to the world but access it through names. One of the benefits of this particular analogy is reinforces the that its juxtaposition of gesture with spoken language of the nomen to the object indicated as strictly deictic, just relationship with the as the word took on iconic qualities through its juxtaposition we might ask ourselves But more importantly, simulacrum. why the ostensibly successful gestures of children constitute only a dry run before these children come into their adult capabilities, why these little index are like the phantom horns of a baby calf. Recall how touch fingers as soon as there is establishes the truth of vision and yet is superseded it possible to "touch" what is not within the grasp of light, light making your hands. Gesture is propped up against the objects it indicates; it only makes sense in their presence. Speech is more useful because the capac? ity of the tongue to recreate the diversity of the world removes the prop: words provoke their own simulacra in the mind of the listener. Although Lucretius says nothing of this here, the shift from gesture to speech with an infinite bank of simulacra intimates a language that operates rather than with the immediate, sensible world, thereby permitting words to disengage from that world.66 in words as naturally deictic, it is not a Given Lucretius' investment that he, like other later Epicureans, drops the later stages surprise of reason and choice in which the application described by Epicurus in the use of names (Ep. Hdt. 75-76). At the same time, our intervenes

of objects provoking us to make noise. The problem is circumvented by following Long and Sedley (1987, 1.64-65) in granting the Epicureans a theory of evolution that allows the inheritance of acquired characters. The child, like the calf, inherits as instincts behaviors developed by the parents. See also Campbell (2003) ad 1011-27. However, I think Lucretius is less concerned about internal consistency here and more interested in aligning naming with deixis. 66See Konstan 1973,48-51. Pigeaud notes that, in the beginning, "varietas n'est pas c'est deja de la nature organisee, distribuee, et non le miroitement du dangereuse, puisque n'importe quoi" (1983, 143). This is what language, however, slides away from. See also Campbell 2003,16-18.

560

BROOKE HOLMES

that the word reading of the account of hearing has already suggested differs in a whole tangle of ways from the visual simulacrum. For Epicurus, sound, whether it is speech or a drum roll, is a "squeezing out" (ekthlipsis) of particles from the body, much as nomina rerum are there to be pressed out by utilitas in 5.1029. Yet in Book 4, it is only the primordia vocum that from the body's depths and given over to articulation are expressed by the lips and the tongue, "fashioner of words." This work at the interface of the world and the body is less instinctual deixis and more creation, less of the world and more of the sound-producing body, than Lucretius And if this ability to create is decidedly human, however much implies. its origins are buried under the operations of impersonal force, then it is vulnerable other technoloto the same forms of excess that undermine gies that start out useful in Book 5.

IV. USES

AND

ABUSES

OF LANGUAGE

In Book 4, the word functioned not only as a form analogous to the simulacrum but also as a created object. But to what extent do articulare and figurare imply atomic manipulation? Once we have begun to inquire into how words are made, to what extent does the line separating letters from atoms in the analogy hold? We know, on the one hand, that the raw material that the mouth deals with does not consist of atoms but rather In of sound particles, which come in a limited number of configurations. this respect, the labor of the tongue and the lips resembles the creation of the perceptible world from atoms, or rather seeds, that is undertaken by the daedala tellus?1 Yet once those sound particles are put into the service of word formation, their qualities cease to matter. Nor does their proper configuration place any particular limit on what can be created. and joined together by the lips, these sound particles behave like Shaped each phoneme elements, serving as the building block of a word. Thus, when sound is formed into a word, one reality is superimposed onto another.68 The word is capable of its own effluences, which take precedence over the communication of the qualities of the voice. This act of formation distinguishes speech, in its advanced stage, then, from its primitive origins, where the phantasia or the pathe shaped the raw material of

67 Bailey describes the seed as "a complex of atoms of such shape and placed in such arrangements that they are now ready to create particular living or inorganic things" (1928, 344). 68It is this collapse of the levels in a Lucretian analogy that Schiesaro 1990 demon? strates.

CRAFTED SPEECH IN DE RERUM NATURA

561

sound. At this advanced level, the automatic imprint is delayed until the vox replicates, and it repeats the contours not of a natural object but of one that has been created. The question of whether to something a word refers accurately real, or, more complexly, whether the real has been correctly reproduced at the level of a proposition or a poem is not raised in Book 4. The decision to place this account of language within the order of sensory is not, in fact, causes the reader to forget that Lucretius experience dealing with a verifiable material world and that, in advanced language, the construction of linguistic reality is no longer provoked by the impact of objects. Such language threatens to stray from the world of strict between and icons, the world that is proper to the equivalences objects simulacra, and enter into the gray area of (artificial? artful?) reproduction. Yet the only "intentional" subjects in the fourth book are the lips and the daedala lingua. By keeping language at the physiological level, to the automatical retains the mechanistic quality attributed effluences. Nevertheless, from within this physiological account, the ad? daedalus gestures to a space beyond the automatic.69 It anticijective pates another epithet given to the lingua at the moment it is claimed by the catastrophic of the plague of Book 6: animi interpres, "translator in the mind" (6.1149).70 By examining how daedalus is used elsewhere poem, we may shed some light on the workings of these half-personified body parts in their efforts to transform, on the one hand, sound particles the excesses of the mind? into words and, on the other hand, res?and into fashioned sound particles. daedalus may refer to both natural and Tellingly, the adjective In its first occurrence in the celebrated artificial production. hymn to Venus, it describes the marvelous fecundity of the earth, whose generative are like those of the goddess herself, the source of all life capabilities Lucretius

69 Perhaps we can see in the word, in addition to the connotations discussed above, an echo of the Epicurean hermeneuein (Ep. Hdt. 76), a verb that itself, as Mackey notes, "must carry the nuance of deliberate, sophisticated, even stylized usage" (2003,11). Like daedalus, popular etymology (e.g., Plat. Crat.408a-b) could find another figure of artifice in the verb that describes what humans themselves are capable of devising (see DRN 4.835). Hermes is a nomothetes rejected by Diogenes of Oinoanda; see fr. 12 col. iii 4-6 (Smith). The Epicurean picture might also be contrasted to the Stoic one, where it is nature alone that takes credit for the cleverness of the tongue: it is made that way (Cic. MD 2.149: "ad usum autem orationis incredibile est. . . quanta opera machinata natura sit"). 70Cf. Plat. Theaet. 206d: . . . to rcoieiv 8ia 9oovfjc; jiexa xr\v ocutou 8idvoiav euxpavfj te Kai r\ zic,xr\v eic, pnuocTCGV ovouixtgov, coa7cep KaTorcTpovuScopxr\v56^av 8KTD7ioiL)|i?VOV bm XOX) OTOUXXTOq pofjV.

562

BROOKE HOLMES

and the poet's Muse (1.21-30).71 The adjective recurs with this (1.1-20) sense at 1.228 and again in Book 5, where the subject is natura: at variae crescunt pecudes armenta feraeque nec crepitacillis opus est nec cuiquam adhibendast almae nutricis blanda atque infracta loquela nec varias quaerunt vestis pro tempore caeli, denique non armis opus est, non moenibus altis, qui sua tutentur, quando omnibus omnia large tellus ipsa parit naturaque daedala rerum.

(5.228-34)

But the various flocks grow, and the herds and the wild animals, and they have no need of rattles, nor must they hear the gentle and broken speech of the foster nurse, nor do they seek different clothes for each change of weather. And most of all they have no need of weapons nor of high walls to guard their own, since the earth itself generously supplies everything for all of them, as does nature, artificer of things. construction at 4.551, Naturaque daedala rerum mimics the grammatical where daedalus takes the genitive of the thing created and modifies the creator (the lingua).12 This is striking in light of the tension in Lucretius between natural and human production. Indeed, it is in this passage that the productivity of natura is highlighted as problematic for humans. Stoic teleology, is cataloguing the earth's Lucretius, perhaps targeting These flaws, which a truly perfect design should preclude (cf. 2.180-81). faults are specifically the ways in which the earth falls short of meeting human needs: large swathes are uninhabitable, it harbors wild beasts that 71"te, dea, te fugiunt venti, te nubila caeli / adventumque tuum, tibi suavis daedala tellus / summittit flores, tibi rident aequora ponti / placatumque nitet diffuso lumine caelum" (1.6-9). It is not clear here whether the adjective is being used in an active sense (wonder-working earth) or a passive one (wonderfully-worked earth), although the em? phasis on generation implies the former. On natura as an active subject, see Kennedy 2002, 90-91. This is the sense in which it is applied by Virgil to Circe: "absenti Aeneae currum geminosque iugalis / semine ab aetherio spirantis naribus ignem, / illorum de gente patri quos daedala Circe / supposita de matre nothos furata creavit" (Aen. 7.280-83). Servius glosses daedala as ingeniosa, since Circe, forging mortal and immortal blood, cleverly fashions a hybrid creature. This Circean quality will become the property of human production. Daedalus was a word with strong epic connotations, descriptive of fantastic craftsmanship and things wonderful to see. It carries a sense, too, of deceptive creation. In Hesiod, for example, daidal- words are always linked to Pandora. More neutrally, Morris points out that "the figure of Daidalos (Daedalus) held special appeal for Latin authors as a symbol of Greek art and as an artist who migrated to Italy in legend, and was popular in art in Italy since the fifth century" (1992, 68). 72See also rerum natura creatrix (1.629; 2.1117; 5.1362).

CRAFTED SPEECH IN DE RERUM NATURA

563

to the prey on man, and, finally, it does not offer enough protection human child (5.200-234). Lucretius' ambivalence is clear. For, despite interest in nature's flaws, the fault seems to lie, rather, with professed human weakness. Unlike animals, these creatures require nurses, clothof intimations fortresses, and, perhaps not accidentally, ing, weapons, of the development a list that foreshadows speech (infracta loquela), culture later in Book 5. Here, the wealth yielded by daedala natura is more than sufficient to satisfy the needs of earth's other inhabitants. Yet nature's very inability to satisfy human needs will open up the possibility and supplant the earth's of artificial production, designed to supplement The adjective daedalus is apwith technological innovation. fecundity its transfer to natura at precisely the point Lucretius foreshadows plied to human ingenuity.73 acme of In Book 5, the adjective at the bittersweet reappears human evolution, the deliciae created by these increas? where, among statues (daedala creatures, we find cleverly fashioned ingly imaginative at and clothing The fortifications, anticipated signa, 5.1451). weapons, hard won by human 5.228 surface here among the bounty (praemia) labor. The statues, however, take their place amid the luxuries, as do the iuvenum simulacra, of Book 2 (aurea... golden statues in the prooemium which are dismissed for being in excess of any need. The word 2.24), these objects' appearing in both cases, deliciae (2.22; 5.1450), designates uselessness. Their creators, like those of the picturae and the carmina named in the same line, ingeniously rework the forms of natura in the of these luxuries colors raw material that it provides. The non-necessity refers to the "highest the book's ominous last line in which Lucretius of the arts" (artibus ad summum . . . cacumen, 5.1457).74 pinnacle 73Cf. 5.811-20, 937-44. In both instances, Lucretius stresses that what natura provides is sufficient. Yet the former passage describes the birth of the human race from earthwombs, a situation superseded by the development of normal means of reproduction. Also, this stage corresponds to the fecund youth of the earth before she grows exhausted and ceases to provide as before (5.826-27). In the latter case, the first humans eat what the earth provides, but they are mauled by beasts and still liable to starve to death (5.1007-8). 74Blickman rightly notes the "delicate ambiguity" of these lines, which, he argues, recall the struggles of 5.1120-42 (1989,186-87); it looks forward, too, to the demise of all the arts when faced with the plague in Book 6. But that technology first responds to real needs is suggested by 6.9-11. For the tension more generally between the benefits of human ingenuity and its excesses in Book 5, see Konstan 1973, 35-58; Asmis 1996. On the non-necessity of music and, by implication, poetry, established already in Democritus (DK68 B144), see Armstrong 1995, 213-15. As Blickman observes, carmina occupy a complicated place in the pre-history for, while the plastic arts receive no respect, Lucretius recognizes in music and poetry a form of pleasure that accords well with his presentation of a rustic ideal; see Buchheit 1984.

564

BROOKE HOLMES

in Book 2, daedalus to the arts to Already appears in relation describe the songs of Apollo's lyre in the context of Lucretius' argument against the possibility of limitless creation, a possibility foreclosed by the finite shapes of the atoms. Read against the leitmotif of unbounded human desire, the passage is fascinating, not least of all in its classification of the lyre's song with the swan's among the world's natural wonders. If the range of atomic shape were unlimited, Lucretius reasons, there would be no check on what could be created. The splendors of the daedala earth?a brilliant color, the taste of honey, Apollo's peacock's be surpassed by ever more marvelous creations (2.500It is precisely the belief in unbounded 506).75 creativity that drives the of utilitas to excess in Book 5. Humans fool themselves into technologies that they can create something in addition to what exists in believing nature. In such a context, it appears likely that there is a correlation between empty desires and the things created to respond to and sustain them. If such things cannot be said to add anything real to the world, it follows that they simply instantiate the excesses of false belief, the surthat Lucretius says the mind adds to the world given by the senses.76 plus In this respect, they appear to function like what Epicurus called "empty sounds" (phonai kenai, Sent. 37; cf. 4.511, verborum copia cassa; 5.909, in hoc uno novitatis nomine inani), although note that Apollo's lyric song the summit of natural creation. represents No doubt the phrase verborum daedala lingua refers, as Mieke Koenen has suggested, to "the rich fertility of the tongue in creating all sorts of words and phrases" (1999a, 457), a connotation that recalls the of nature in forming a range of compound ob? generative capabilities But might it also refer to the tongue's creativity in a broader sense, jects. namely its capacity to fashion words to provoke images that might stand in for the world of present objects? We may go further still to point out that, as De rerum natura itself evinces, the daedala lingua is also able to a real world that may only be "seen" analogically or through its effects, that is, the world of atoms and void. The agility of the perceptible summon carmina?would

75The sense is clear, but the passage is corrupt. Munro prints a lacuna after 501. Giusanni detected another after 499. Lucretius often returns to the theme of natural limits: see 1.199-204, 551-98, 746-48; 2.496-521, 718-19, 1120-22. Most pertinent here is the speech of Natura at 3.931-49, where the fear of death is expressly related to unlimited desires; see also 3.1076-94; 5.168-73 (the gods have no desire for novelty), 1430-33. On the relationship between irrational fear (that is, fear without cause or anxiety), especially the fear of death, and limitless desire, see Konstan 1973; Deleuze 1990, 272-73. 76SeeKonstan 1973,49-51.

CRAFTED SPEECH IN DE RERUM NATURA

565

the diversity of objects in the world. tongue succeeds first in mimicking Yet it is clear that Lucretius is well aware that language is not merely a of things in the world, to be performed description equally well by different speakers, who might err only in the "version" of reality they the first names are spoken engage, as Glidden argues (1983). Although under compulsion, in the later stages of the Epicurean account people choose their words with the aid of reason.77 Deliberate a level of mastery over the body's speech introduces innate capabilities. Given that the first sounds emitted imitate nature not the conditions for the creation by copying its forms but by reproducing of difference, such mastery opens the door to a language independent of the natural world. The space traversed by the act of choice destabilizes the "translation machine" and its truth claims: choice turns the tongue into an interpres, not of the world but of the mind. As such, it becomes ideally somewhere the excesses, capable?somehow, offstage?of incarnating both the false opinions about the gods and the true beliefs about microphysical reality, which the mind adds to what it suffers of the world, thereby breaking the direct line between object, pathe, and sound. Thus, the iconoclastic interlude between my prolepseis and yours is subject to all kinds of additions to which the mind is prone, sometimes rightly, often as the prooemium to Book 6 suggests:78 "[Epicurus] understood wrongly, that it was the vessel itself that produced the disease (vitium) and that by this disease all things were corrupted within it, whatever came into it outside (my italics)... ,"79 It is under such conditions that gathered from

77 xoix; [jiev otiv] avayKaadevxaq avoccpcovfjooci, 8e xa>A-oyiofiw xoix; e^ojjivcnx;,mxa xfivnXeicxr\vaixiav ouxccx; epuriveuoai (Ep. Hdt. 76 = 19A Long and Sedley). On dvoupoovdv, see Mackey 2003, 7-8; Sedley 1973, 59. 78Verbs that stress the addition of opinion are common in accounts of the origins of religion and the gods: see 4.1183 (tribuisse); 5.164 (addere), 1172 (tribuebant), 1175 (dabani), 1195 (tribuit, adiunxit). This addition of something must account for "the rhetorical effects of language that are not evident at the level of what is, in reductionist analysis, isolated as the individual word or phrase" (Kennedy, 2002, 88). 79 "intellegit ibi vitium vas efficere ipsum / omniaque illius vitio corrumpier intus / quae collata foris et commoda cumque venirent. . . quod taetro quasi conspurcare sapore / omnia cernebat, quaecumque receperat, intus" (6.17-19, 22-23). Sedley observes that for Epicurus, in the context of advanced language, "to apply a name to an object is to express an opinion" (1973,19); he cites On Nature 28 fr. 6 col. i 5-13; fr. 8 col. iv 4-9; fr. 11 col. ii 510; fr. 13 ii 4-2 inf.; fr. 6 col. ii inf.-7.13 sup. As Wigodsky writes, "it would in fact be very surprising if [Epicurus] had envisioned a purely cognitive use of language, since his denial of an intermediate condition between pleasure and pain means that there is no room in his psychology for a cognitive state untinged by emotion" (1995,62). Equally surprising would be a purely behavioral model, in which the mind in no way intervenes in speech production.

566

BROOKE HOLMES

the paradox of the "empty word" becomes possible. Yet, as was sugat 4.386, this vitium is like a pharmakon, not only the gested designating of false belief but also the panacea, i.e., true ratio (= reason), the poison second sight capable of puncturing the kaleidoscope of Lucretian imag? and penetrating to the heart of the atomic real. Vitium captures the ery underside of daedalic capability while hinting at its promise. It has long been recognized that one of the abiding concerns of was the correct use of language. In a well-known from Epicurus passage the beginning of the Letter to Herodotus, he urges his reader to grasp "what underlies words" (hypotetagmena tois phthongois, Ep. Hdt. 37), described there as the proton ennoema.80 However we may understand this phrase, it is clear that it was crucial to Epicurus to reestablish the bedrock of language, that which is rooted in the sensible experience of the world.81 The danger identified by such a caveat is that in believing and and wondering, "everything will be undistinguished investigating (akrita) us as we engage in an endless chain of proofs or we will have empty by words."82 Given that this passage occurs in the beginning of the Letter to a precis of Epicurus' physical doctrines, perhaps it should be Herodotus, evaluated with its context in mind, that is, as a key to reading the Letter itself. For the threat of outstripping the limits of what exists at the microor the macrophysical level looms large precisely because philosophizing about atoms and void necessarily operates at one remove from the sen? sible. What Epicurus may be alluding to here is what we might call nominal glide, a process adumbrated in the third stage of his account of the origins of language, when "people sharing knowledge, introducing certain unperceived entities, handed down certain utterances" (tivoc 8e Kai ou auvopcojieva 7ipdyjiaxa eiacpepovxaq xovq croveiSoxac, napEjyvr\aai here should be xivaq (pGoyyouc,, Ep. Hdt. 76). That "certain utterances" words applied to non-sensible via analogy entities, understood existing with the perceptible world (e.g., to kenon, "void"),83 is an interpretation that gains support from a recently reedited fragment of Philodemus on

80See above, n. 44. 81On Nature makes clear that the mature Epicurus had decided that the proper use of language would draw on everyday language rather than attempt to recover a primitive model or start anew based on sensory observation. This allies him, then, with linguistic conventionalists; see e.g., D.L. 10.13. See also Sedley 1973; Asmis 1984, 34. 82Cf. Sent. 37; D.L. 10.34 eivou xaq fiev Txepi Ttpayudxcov, 8e xcov (xcovxe ^nxriaecov xaq xfiv nepi \|/iX,r|v cpcovriv). 83Glidden argues they are coined (1983,205). Long and Sedley see here, rather, the application of existing words (1987,1.100); see also Sedley 1973,16; Mackey 2003, 9-10.

CRAFTED SPEECH IN DE RERUM NATURA the origins of poetic language Poems 5 by David Armstrong, to appear in a forthcoming edition Fish, and J. I. Porter:84 Jeffrey

567 of On

Ka0]6ta)\) xe yap f] 7ior]l[TiK]fi^evo9cov{a 7ial[pfi^]0ev ziq avGpcorcoix;I [5ia] tftXoxvniav xcov I [eiq x]aq Koctvaq cocpE^Clfaq kou]vco<;xatq ?pur|l[v?ioci<;] *85 Ke%pr]|ievcov (PHerc 403 fr. 5 col. i 8-14 Armstrong and Fish = fr. 6.8-14 Sbordone) For, in general, the strange language of poetry came to mankind through emulation of people employing expressions in new ways for new benefits. (trans. Mackey) While PHerc. 403 fr. 5 col. i is framed by two long segments (50-55 lines) of missing or illegible text, the recoverable context suggests that it is part of a discussion on the Stoic practice of allegorical reading, which sought to discover the hidden wisdom of a poem.86 It most likely represents Philodemus' own argument on the origins of poetry's figurative language in new ways (Mackey, 2003, 23-25). The "people employing expressions

841print the text from the forthcoming edition of Armstrong, Fish, and Porter with their kind permission. I have omitted lines 1-7 and 15-17, which are too fragmentary for secure reconstruction. Recent work on the fragment, aided by the use of multispectral photography, has succeeded in identifying misleading errors in the editions of Jensen (1923) and Sbordone (1971). Of particular import for the text's reconstruction and its relevance to the above arguments is Armstrong's recent restoration of the rare noun of ^evocpcGvCoc, "strange language," for the accusative Sevocpcovxa previous editions. A thorough and nuanced assessment of the fragment's import for Epicurean theories of language was undertaken by Jacob Mackey in an unpublished Oxford M.St. thesis (2003); my observations above are greatly indebted to his work. This is the earliest attestation oixenophonia, although Mackey (2003,19-20) cites an instance of the word from the De tropis, attributed to the grammarian Tryphon, a rough contemporary of Philodemus. Tryphon describes a schema as an "intentional error (hamartemahekousion, as opposed to a hamartema akousion or grammatical solecism) made by a poet or prose-writer through art or xenophonia or literary ornamentation" (26.1 West). 858 Ka9]6Xo\) Janko. 8-9 te 7iotiItik]tiMackey. 9 ^evocpcovloc primum agnovit hic recte contra lectionem et papyri et exempp. Nap. ("disegni") alii legi Gomperz, Hevocpcovxa editores adhuc. 9-10 7ialpf|X]9evArmstrong, Janko. 11 8ia Armstrong. 12 ei<;Jensen. x]a<; Armstrong. 12-13 cb(peM[a<; Armstrong. 13 kou]vco<; Armstrong. 13-14 epixrdyelaic]Jensen. Supplements from Janko were suggested to Armstrong, Fish, and Porter privatim; for Jensen's, see Jensen 1923; for Mackey's, see Mackey 2003. 86PHerc 407 begins the discussion of the "good poet"; see the fragments published in Mangoni 1992. On the context, see Mackey 2003,17-18, and, on the critique laid out in On Poems 5 more generally, see Asmis 1995b. On the tradition of allegorical reading in antiquity, see the overview in Gale 1994,19-26. For Lucretius' own mockery of allegory, see 1.641-44.

568 for new benefits" echoes

BROOKE HOLMES the reference

to the people who "sharing knowl? entities, handed down certain utedge, introducing unperceived terances" in Epicurus' third stage. The similarity of the two phrases lends that Epicurus admitted a useful stage of nomisupport to the hypothesis nal glide, during which people like natural philosophers, who managed to "see" the workings of the natural world at a subphenomenal level via certain reasoning, described these workings by transferring words that are proper to phenomenal to microphysical experience reality.87 This particular means of expanding the referential field of various words not only enables people to share knowledge but also, one might world firmly tethered to the suspect, keeps insight into the microphysical of the macrophysical world, as Epicurus' caveat on language experience reference to what one knows via the recognizes. By making continual senses, one has the best hope of keeping language from spiraling off into interminable proofs and proliferating "empty words," words, we may imagine, that fail to refer to any reality at all, macro- or microphysical, and externalize of the mind, a false belief (hypoonly a false movement is particularly lepsis pseudes, Ep. Men. 124).88This lack of discrimination when one is dealing with what cannot be directly confirmed dangerous by the senses (the gods, the atoms), a situation that has much in common with a state of sleep (4.757-64). One risks mistaking, as Deleuze says, the false infinite for the true infinite (1990, 277-79). What may be at stake, then, is the fragile legitimacy of usefulness captured above by Lucretius' daedalus and its gesture to the point at which ingenuity threatens to use value. If it is through the peculiarly human skill of epilogismos outstrip that one perceives the workings of nature, thereby extending the overof touch by sight a step further, the condition of this action is also taking the space of error, excess, limitlessness.89 Lucretius does refer to the perils of nominal glide in Book 3, albeit within the bounds of philosophical when he rejects the use speculation,

87See Mackey 2003, 20-23. That Epicurus has natural philosophers in mind here is argued by Sedley (1973, 19). Mackey observes that the verb synoran in Epicurus consistently has the sense of mental "seeing" or inference (2003,10); see also his discussion at pp. 24-28. See also On Nature 28 fr. 8 col. iv. 1-9 (Sedley), referring to the "first man to think of void in terms of immediacy and time and place." On the difficulty of naming void, see Porter 2003, 201-8. 88See Everson 1994a, 103-5. 89See 2.1044^17, on imagining other worlds: "quaerit enim rationem animus, eum summa loci sit / infinita foris haec extra moenia mundi, / quid sit ibi porro quo prospicere usque velit mens / atque animi iactus liber quo pervolet ipse."

CRAFTED SPEECH IN DE RERUM NATURA of the word harmonia anima to the body: to designate the relationship of the animus

569 and the

quapropter quoniam est animi natura reperta atque animae quasi pars hominis, redde harmoniai nomen, ad organicos alto delatum Heliconi; sive aliunde ipsi porro traxere et in illam transtulerunt, proprio quae tum res nomine egebat. quidquid id est, habeant: tu cetera percipe dicta.

(3.130-35)

Hence, since the nature of mind and soul has been shown to be in some way a part of man, give back the name of harmony, handed down to the musicians from high Helicon: or else they themselves dragged it forth from somewhere, and transferred it over to this thing, which at that time was lacking a name of its own. Whatever it is, let them keep it: you look to what else I have to say. The forced of the word from where it once belonged and its dislodgment the musicians (aliunde ... traxere; transtulerunt) antici? appropriation by a second, more insidious transfer of the word into the philosophical pates His reference to a sphere, from which Lucretius banishes it definitively. which a word is "dragged over" to describe a res that lacks its process by own name confirms the hypothesis that imperceptible things were prima? named through catachresis. His lack of concern for how naming rily takes place in music (quidquid id est, habeant) stands in direct contrast to his policing of the philosophical vocabulary. For the improper extension of harmonia to the relation of mind, soul, and body is dangerous. By assimilating it to something that it does not, in fact, resemble, as Lucretius these Greeks have drawn the wrong attempts to demonstrate (3.106-29), conclusion, namely that the animi sensus is not located anywhere in the body.90 In doing so, they have wandered far astray, an error of metaphorical perversion now embedded in the word itself.91

90See On Nature 28 fr. 8 col. v 6-7 ov (Sedley): |iex?0?U?0aI cruvi86vxe<; [xoia]uxa ek IxxvoqeniX[oy]iG\i[o]x): altered [certain names] when by some act of empirical reason"we ing we saw that they were not of this kind" (trans. Sedley). Lucretius extends this process to more complex metaphors as we see above with harmonia; see also 3.359-69 against understanding the eyes as the "doors to the mind." 91 magno opere in quo mi diversi errare videntur, 3.105. This notion of erring recurs often in Lucretius (e.g. 1.846; 2.10,82,740; 4.823; 6.67); on path imagery, see West 1969,7274; Volk 2002, 89-91. For perversa ratio, see e.g. 4.833; caeca ratio, 6.67.

570

BROOKE HOLMES

This talk of "dragging" words around confirms the existence of a of choice in speech that is absent from the stage of deictic nominaspace tion. While there may be some philosophical benefit to be gained from the elasticity of words, Epicurus also claims in On Nature 28 that all error has no other form than the one related to prolepseis and phainomena "on account of the many-turning habits of speech" (8ioc xovq TroAAnporcouc; efOiJojmnx; xcov Ae^ecov).92 Error seems to arise from our misinterpretaon account tions of the data provided by our prolepseis and phaenomena of the entrenched of language. Yet this polytropy polytropy emerges from a dilation of the referential field that is also useful, eis tas kainas Philodemus observes that all techne ophelias, Philodemus says. Elsewhere "could not utter a word if deprived of the utility of metaphors."93 There is evidence suggesting that Epicurus, too, recognized metaphor as a mixed As for poetic but also necessary to philosophy.94 blessing: dangerous, it as a trope built onto a Philodemus seems to understand language, of a word's potential for infinite displacetrope, useless and symptomatic ment. Whether Lucretius would follow him on this is unclear. It is true of carmina, a word denoting both music and treatment his sensitivity to their pleasures while the poetry, betrays (5.1379-1411), of carmina among the luxuries at the close of Book 5 may inclusion concede their uselessness.95 And while he says nothing explicit about 92On Nature 28 fr. 12 col. iii 6-12 (Sedley): naaa r\ aujajpxia eaxiv I xcovdv9pco7rcov I ovbkv k'xelpov axfjua r\ xo erciIxcofi7ipo^f|\|/ecov yiyv[6] lu?vov Kai xcoficpaiv[ou]evcov e'xo-oaa Sia xauq noXvxponovc, e[0i]la|io\)(; xcov>i^ecov .... See Long 1971, 123. Cf. Glidden 1983, 219-24. 93Phld. On Rhetoric 4 col. xv 15-18 I ox> (1.175 Sudhaus): 7iaaa xe%vr|q>cov[r|]v 5-uvaxai Cf. 7cpo[iea]6ai axep[r|]6eTaaxr\q?k xcovI (xexacpopcov ex>xpr\o-[xiaq]. id. col. xxi 8-15 (1.180 Sudhaus) and Wigodsky 1995, 62-64. 94See On Nature 28 fr. 13 col. v inf. 3-6 (Sedley) on "transferences of words from the class of the known to the unknown" (... aX\Xo[xx;.] erc! 17ioi[eiv<pcovco]v xa luexacpopac; The ayvcol[axa]v[nb xcoy yv]coaxcov). context seems to imply that the transfer in itself is not wrong but that those making it have erred; see Sedley 1973,62-65, and Mackey 2003,25-26. 951find attractive Armstrong's reading of vigiles in 5.1408 as the Hellenistic heirs to lyric poetry rather than watchmen, as most editors and translators do (1995,214, n. 10). As he notes, the account that follows of how rustic staples are given up for civilized luxuries implies that these poets offer a more complicated version of early song that in no way delivers more pleasure than the old one. Buchheit 1984,156-58, proposes that Lucretius sees himself as an heir to these pastoral poets, whom he understands as figures from a Golden Age. However, the complications of reading any kind of Golden Age in Book 5 speak against any easy association between Lucretius and the past; he is as likely to identify with the vigiles. Else where in Book 5, the poets' songs serve to hand down past events, such as the Trojan War or political history (5.1444-45). The mention at 5.1445 of the discovery of the alphabet (nec multo priu' sunt elementa repertd) appears to be tied more closely to the following two that his own

CRAFTED SPEECH IN DE RERUM NATURA

571

linguistic innovation unique to poetry, he is highly sensitive to the perils of metaphor and language in general, not only among the poets of but among philosophers, too. mythology Despite cutting short his account of the origins of language with humans using their arsenal of sounds to mark off the world's diversity, Lucretius does gesture towards the afterlife of these deictic exercises in the very next section. Having accounted for the discovery of fire (5.1090104), he moves on to the shift from ingenuity to greed, as the intelligent of fire give way to the rise of cities and kings, the discovery applications of gold, and the invention of property. Surveying this dystopia, Lucretius denounces with characteristic verve the ambitions of those striving to acquire wealth and avoid death, and he blames this ambition precisely a disengagement from what is known via the senses and a privileging what is heard from the mouths of others: proinde sine incassum defessi sanguine sudent, angustum per iter luctantes ambitionis; quandoquidem sapiunt alieno ex ore petuntque res ex auditis potius quam sensibus ipsis, nec magis id nunc est neque erit mox quam fuit ante. on of

(5.1131-35)

So let them sweat blood, worn down in vain, struggling along a narrow path of ambition, seeing that what they know is from the mouths of others and that they seek things in accordance with what they hear, instead of in accordance with their own senses. And it profits no more now?and never will?than it did before. Though recently in a world of happy deixis, we suddenly find ourselves before words that circulate with reckless autonomy, their message diametrically opposed to what is delivered by the data of the senses. These words arise alieno ex ore, from a mouth that not only belongs to someone else but is also estranged from the real world. In this world full of empty hearsay, it seems that the daedala lingua is a machine gone berserk. with the The formation of a world in words capable of competing atoms and letters is world confirms that the analogy between physical

lines ("propterea quid sit prius actum respicere aetas / nostra nequit, nisi qua ratio vestigia monstrat"), which echo the Thucydidean archeology, than to a theory of civilized poetry as "the play of written letters, elementa, on the page," as Armstrong argues (1995,215). At the same time, the discovery of the alphabet occurs at a point where innovation is at its most perilously ingenious (immediately after the rise of civilization, with its excess desires, has stirred up the tide of war, belli magnos commovit funditus aestus, 5.1435).

572

BROOKE HOLMES

only an analogy, but language is all the more dangerous as a result. Words both are and are not things. Lucretius goes to great lengths to prove that voice is corporeal because it does things (Koenen 1999a, 445-48), but we are more in the world of J. L. Austin here than the world of body and void. Once voice is fashioned into a word, it remains capable of impacting the about the senses of another person. Yet it no longer bears information Rather it conveys whatever the speaker wants to body. sound-producing say, staging in the mind of the listener not only a particular vision but also of a the speaker's beliefs. The fashioning of a sentence is the fashioning There may world that feeds on matter even as it skews its representation. be constraints on what can be said that are imposed by the internal confines of a given language, but once words and sentences and poems are created, the immutable laws of nature do not apply. Only ratio, in tandem with a testing against the senses, can enforce these laws at the level of the said. The problem with operating at the level of shared reality the via words is that someone may have gotten it wrong, contaminating entire referential Listeners come to mistake words for frageconomy.96 fiction.97 ments of the world, rather than components of a manufactured is a sense excluded from the true senses to the extent that Thus, hearing it is reduced, as it was in Book 4, to the reception of human speech and its "pile of empty words," infected with the vitium of the mind. a similar anxiety The excursus on Cybele in Book 2 demonstrates about contamination and the damage done by words. But it also admits own gambles of complications and complicities that hint at Lucretius' of the daedala lingua is in full with language.98 The creative activity evidence as Lucretius has the old Greek poets yoke beasts to the Earth Mother's chariot and wreathe her head with crowns, stories that provoke, in turn, the creation of an imago of the goddess that mimes not natura but lingua (2.609). He recognizes, as Philodemus does,99 that things set forth very well (bene et eximie, 644) may still be estranged from the truth (longe sunt tamen a vera ratione repulsa, 645). Having set Memmius 96Memorable here is the linguistic ingenuity of the lover (4.1153-70), although this is no doubt benign compared to the things that people say about gods. 97For the use offingere with false stories, see 1.104, 371, 643-44,1083; 2.58; 4.581; 5.164 (adfingere). 982.600-660; cf. 5.405; 6.754. On this passage, see the discussion in Schrijvers 1970, 50-59; Gale 1994, 26-32. As Gale points out, Lucretius does not endorse the systematic allegoresis of the details of the cult, which is as guilty of propagating false belief as the rites themselves; cf. Bailey 1947, 2.898-901. 99See, for example, On Poems 5, PHerc 1425 col. 25 30-col. 26 20 (Mangoni); id. col. 32 3-22 (Mangoni). For Philodemus, the only poetic good is the creation of a resemblance, although this has no moral utility. See Armstrong 1995,215-25; Asmis 1995b.

CRAFTED SPEECH IN DE RERUM NATURA straight on the gods?that removed from ours?he they live, placid concludes: and immortal,

573 in a world far-

hic siquis mare Neptunum Cereremque vocare constituet fruges et Bacchi nomine abuti mavult quam laticis proprium proferre vocamen, concedamus ut hic terrarum dictitet orbem esse deum matrem, dum vera re tamen ipse religione animum turpi contingere parcat.

(2.655-60)

If anyone is determined to call the sea Neptune and corn Ceres, and prefers to use the name Bacchus than to produce the true name of the juice, let us grant that he may proclaim that the world is the Mother of the gods, as long as he, by means of the truth, keeps himself from staining the mind with shameful religion. Nominal glide resurfaces here in the form of metonymy, that is, using the name Bacchus instead of the proprium and Lucretius is well vocamen, aware of the risk it carries. Note the verb abutor, which can mean both "to make full use of" and "to exploit, abuse, misuse [language]" (OLD, sv. abutor). It only recurs at 5.1033, where, we may recall, it describes lan? of the tongue's capacity for markguage as humans' natural exploitation ing difference (sentit enim vis quisque suas quoadpossit abuti). Designatuse of one's faculties and the potentially ing both the opportunistic reckless abuse of names, the verb occupies the same charged zone as daedalus. The dum ... tamen construction weighs heavily on the decision to use metonymy. Its detailing of the strict conditions under which this a high risk of failure. The havoc wreaked establishes usage is permissible of the poets' insight, beelaboration by myth, which is the illegitimate of the consequences comes evidence of this failure.100 100 That the poets' insight is derived from an essentially correct prolepsis is also suggested by Konstan (1973,25, n. 59) and Gale (1994,32, n. 114). Faulty beliefs about the gods are always said to arise from incorrect inference: see e.g., 1.151-54; 2.167-81,1090-92; 5.76-90, 114-45, 1186-1240; 6.48-91, 379-422, 760-68, and Gale 1994, 130-38. See also 3.624-30, where Lucretius says that we are incapable of imagining an immortal soul that is not endowed with sense, perhaps because this faulty belief is based not on sheer fantasy but on a real image?the simulacrum of a dead person (4.760-61)?from which we have made a wrong inference. Nor can the painters or poets imagine it otherwise. At the same time, something more complex is going on here with respect to the limits placed on our imagination by our beliefs. Porter observes of the attempt to imagine our own deaths (3.870-93), "the question as to how we can represent death to ourselves becomes a question as to how we can detach ourselves from our conventional ideas about the body, given that our conventional ideas about ourselves are so intimately wrapped up in our view of ourselves as embodied souls" (2003,202). On this passage, see also Holmes (forthcoming).

574

BROOKE HOLMES

that one is guided by true reason (vera Nevertheless, provided one may call the earth "mother." For it is true that the poets' re),m with the generative insight is consistent capacity of the earth. And four hundred lines later Lucretius himself is calling her "mother," evidence that the use of metaphor is a risk that he is willing to take in the interest of lending light to the obscure world of atoms and void.102 Lucretius' to improper uses of metaphor sensitivity may arise from his own commitment to metaphorical which builds on the figures of lan? language, guage that he inherits from Epicurus (e.g., void; see 1.426), as a means of uses making atomic reality visible. For it would seem that Lucretius and not as a as a tool to philosophical figurative language exposition trope built onto a trope, as in Philodemus. Despite the potential pitfalls, is indispensable. That this is so betrays a more basic truth for metaphor that language is necessary to make people see even Lucretius, namely what is before their eyes, that the words alieno ex ore are a precondition of knowing where to look, what to feel. Thus, Lucretius' project, in one is an attempt to reintroduce the original triangulation of language, sense, where a speaker pointed to objects to which the senses of the listener reference to the to make frequent might also appeal. This tendency of his reader participates in this strategy to keep words from experiences like the simulacra of dreams, impossible to verify,103 even as functioning the true object of investigation hovers out of the senses' reach. At the to same time, even at the level of perceptible things, these exhortations

101 Schrijvers 1970, 58, reads vera re as an instrumental ablative modifying parcat rather than as an adverb; see also Gale 1994, 31, with n. 111 (vera re anticipates the vera ratio of 5.406). The passage, to the extent that it establishes the conditions for the proper use of language, can be seen as analogous to Ep. Hdt. 37-38. Despite the caveat, Lucretius does permit the metonymy, at least for himself, pace Clay 1996, 781-82. 102 quaproptermerito maternumnomen adepta est, 2.998; cf. 5.795-96,821-22. Kennedy offers that "the poem could be seen as playing a dangerous rhetorical game for high stakes" (2002, 92-93). See also Schrijvers 1970, 57-59. 103 1.102-6, Lucretius makes this At relationship between false speech and dreams explicit: "tutemet a nobis iam quovis tempore vatum / terriloquis victus dictis desciscere quaeres. / quippe etenim quam multa tibi iam fingere possunt / somnia quae vitae rationes vertere possint / fortunasque tuas omnis turbare timore"; cf. 3.1046-52 on the life that is a waking dream. As Gale points out, Lucretius never uses the word fabula in De rerum natura: somnium comes closest to describing the delusions generated by false belief (1994, 26, n. 94). If these stories prey on fear, others seduce by means of pleasure: see 1.641-44; 4.592-94. What Lucretius' words do, rather, is draw attention to the simulacra of true things always before our eyes ("cuius, uti memoro, rei simulacrum et imago / ante oculos semper nobis versatur et instat," 2.112-13, using the motes in a sunbeam to prove the existence of unbound atoms); see Schiesaro 1990, 26-30; id. 1994, 86-87.

CRAFTED SPEECH IN DE RERUM NATURA

575

see this or that phenomenon shape a reality that is, as Duncan Kennedy "the effect of description" (2002,70), quite literally in Lucretius. points out, of seem to form the cornerstone These calculated acts of evocation For Lucretius is quick to stress that his work is, the poem's usefulness. the Latin language for new benefits.104 Referabove all, useful, employing ring to his account of void, he tells Memmius:

quod tibi cognosse in multis erit utile rebus nec sinet errantem dubitare et quaerere semper de summa rerum et nostris diffidere dictis.

(1.331-33]

To know this will be useful to you in many things, and it will keep you from wandering in doubt and from always wondering about the universe and from distrusting my words. And again, following his proofs that the corpora of the mind are exceed? ingly small and mobile, he tells Memmius, "and this truth, when known to you, will in many things, friend, prove useful (utilis), and will be reckoned This usefulness of service" (3.206-7; cf. 3.417-20; 4.25; 5.113; 6.938-41). of novelty at heralds the skill of the craftsman. While the celebration the exploration 1.926-30 (4.1-5), where Lucretius presents his poem as a Roman poet's of what has never before been sung, is, paradoxically, visit to a Callimachean topos,105 it serves also as a reminder that language knows only the limits of its daedalus creator, who constructs a picture of what can only be seen through an act of sustained imagination:106 Nec me animi fallit Graiorum obscura reperta difficile inlustrare Latinis versibus esse, multa novis verbis praesertim eum sit agendum propter egestatem linguae et rerum novitatem; sed tua me virtus tamen et sperata voluptas suavis amicitiae quemvis efferre laborem suadet et inducit noctes vigilare serenas quaerentem dictis quibus et quo carmine demum clara tuae possim praepandere lumina menti, res quibus occultas penitus convisere possis.

(1.136-45)

104 Cic. Fin. 1.71-72, where Cf. Torquatus says that Epicurus would not read the poets "in quibus nulla solida utilitas omnisque puerilis est delectatio." 105 Volk 2002,87, with the See bibliography at n. 52 on Callimachean influence here. 106 the On "poverty" of his native language, see also 1.832; 3.260. On his role as primus, 5.337. See Dionigi 1988,11-14, on the novelty of Lucretian language.

576

BROOKE HOLMES Nor do I fail to see that it is difficult to illuminate the dark discoveries of the Greeks in Latin verse, especially since we must employ new words on account of both the poverty of the language and the newness of these things; but, nevertheless, your merit and the hoped-for joy of your sweet friendship persuade me to undertake this labor. They lead me to stay awake through the calm nights, trying to discover with what words and with what song I might finally lay before your mind the clear light with which you might gaze into the depths of hidden things.

The distance between compulsory naming and the struggle to develop a real could not be greater. to gesture towards an imperceptible language As the hinge between these two stages, the daedala lingua emerges as the tool of a mobile mind, capable of summoning up absent objects in the mind of another person, capable, too, as interpres animi, of communicating that mind's errors and its insights. The craftiness of the tongue points to the poet who puts it to use. For it is the lingua that sends forth from to the ears (per auris) the poet's own breast his rich store of arguments "A new thing" (nova res), by which we might of his listener (1.412-17). the poem itself, "is struggling to fall upon your ears (ad auris understand In this and a new view of things to reveal itself" (2.1024-25). accidere), where the work of the daedala lingua effects a nova species collocation, of the word is exaggerrerum in the listener's mind, the visual dimension in order that Lucretius might claim his words as "traces" (vestigia, ated, to knowl? as a springboard 1.402) that function, like sense perceptions, the similarity of the simulacra that strike the mind to edge.107 Exploiting those that strike the eyes, he restages the deixis of primitive language, in in the which the subject and the world enjoy a natural pas-de-deux context of atoms and void. motif grounded of the poem as revelation?a This representation in the iconicity of the verbal artifact such as it emerged in Books 4 and 5?is a defining feature of De rerum natura, oft noted by critics and goal in emphasized by Lucretius himself. Indeed, the poem's professed simile is to make Memmius see (perspicis) the nature of the honeyed-cup things, the figure in which it is shaped (qua constet compta figura, 1.950). This seeing is imagined as the effect of a process whose construction Of course, should be obliterated by the act of sight (Kennedy, 2002,73). of this process when it suits Lucretius is willing to reveal the mechanics of his purpose, that is, when they may render visible the mechanics

107 the On revelatory quality ofthe poem, see Schrijvers 1970,38-47; Gale 1994,14445; Kennedy 2002, 71-73.

CRAFTED SPEECH IN DE RERUM NATURA atomic

577

anal? reality. For, as the most common version of the letter-atom elucidates, revelation is realized through Lucretius' own manipulation ogy that to create a kaleidoscopic of the elements of language experience and mental insight. That analogy taps the shuttles between imagination to force a simulta? of sight, hearing, and imagination triple resources of words on the page, together neous perception of the elemental reality with the specificity of their sounding and the object. In such a model, the sounds and objects at the between lack of morphological resemblance of the sound by the birth of language is confirmed by the representation in the register of the visual confirms this lack letter, whose reemergence an The letter is thus freed to become of iconic or formal consistency. en? of the atom, whose shifts in order and position generate analogue level. Of course, the sounds tirely new objects at the epiphenomenal were never akin to the constituent elements of the object but arose from in the its interaction with another species of body; they are symptoms of forces. But in advanced sense that they result from an encounter language, sounds are tied to the letter. The attention in Book 4 to the spoken word as a frangible object retraces the elemental nature of the word at the level of sound. In doing fall away it repeats the process by which the means of revelation so, before the effect. As we saw there, words only have meaning qua words. Their properties qua sound are rendered immaterial by the act of vision. The articulated word in such a context is opposed to mere sound, to the in its to disappear extent that its own sensible qualities are supposed that is, in its articulation. imitation of the object's morphology, analogical Thus the spoken word comes to mimic the diversity of the world in a to order, as well as through limited way, primarily through changes each change causing the word to signify combination (1.827; 2.1014), is crucial different. The verb significare 2.1016) something (significare, here. It cues us to think of second-order simulacra, the sea or the sky or the sun, rather than the sound of a sounding. For it is not only the quality of the primordia vocum that ceases to matter in the discussion of speech of the spoken word, i.e., the phoThe constituent elements production. nemes, are also without meaning, as is made clear in Lucretius' meditation on the collapse of speech into sound in Book 4. It should come as a surprise, then, that De rerum natura is a poem teeming with sounds that refuse to yield readily to the image. Having as a means of tracked the assimilation of the word to the simulacrum accessing res ipsae for much of this paper, I would like to close by ceding a place to this obstinacy. For however generous a tolerance of poetry we elicit from our Epicurean evidence, De rerum natura still feels heterodox.

578

BROOKE HOLMES

Lucretius, despite attending to language at the level of the word, seems of a language dividends deeply attuned to the risks and the potential rendered deictic not through the transparency aimed at by Epicurus' cardinal linguistic virtue sapheneia but through the exploitation of the vast resources and complex texture of language itself. I suspect the commitment to the poem's visual and acoustic components, exand implicit, respectively, is less motivated a sense of formal plicit by of the letterof the multiple implications unity and more symptomatic atom analogy with which I began. Lucretius is a poet sensitive to the rich of similarity, difference, and creation that these implications At 1.828-29, having noted the effects of a shift in the order of support.108 letters, Lucretius tells us that the atoms have at their disposal many other means "by which all various things can be produced" (unde queant variae res quaeque creari). But, in fact, phenomenal diversity for him can be reduced to two causes, namely the particular composition of a compound and the different shapes of the atoms themselves. The motif of creation at the level of speech references the first type of cause and its model of blocks whose organization building might produce a reality effect in the listener that is shot through with insight. The word masquerading as simulacrum offers the promise of that medium, the truth of touch with the distance for contemplation. The latter type of cause is discussed by Lucretius from 2.333-729, where, we might recall, the only version of the letter-atom analogy that stresses differences among letters appears (2.688tar99). Particularly intriguing here is that a major class of phenomena this explanatory schema is the variation in sensation itself, geted by 108 Thury 1987, 284-85; Armstrong sees the acoustic texture as conflrmation of Lucretius' Philodemean poetics (1995, 224-28). For Philodemus, elaborate sound patterning could only get in the way of the communication of verbal content; see On Music 4 col. xxvi 9-14 (Kemke); id. col. xxviii 16-35 (Kemke). Cf. DRN 1.643-44, on the fools who read the verses of Heraclitus: "veraque constituunt quae belle tangere possunt / auris et lepido quae sunt fucata sonore." It is here that Lucretius sounds the most like Philodemus and the most like himself at 1.934 (musaeo contingens cuncta lepore, speaking of his own work; cf. 1.15,28 from the address to Venus).The easiest solution is that the fools (stolidi) who praise Heraclitus mistake what sounds nice for truth. This is most likely correct, and I doubt that Lucretius would say that there is anything true or false in the sounds of De rerum natura, which is not to say that he does not recognize his skill as extending to the composition of the poem as a whole. He is certainly ready to grant that the Greek poets have set forth the story of Magna Mater bene et eximie (2.644). The real question is whether the effects of the words qua sound have any independent worth for Lucretius, that is, if there is something that they convey of natura rerum that does not translate into rational insight. I think that, despite the explicitly instrumental role of the poem's charm (which is not necessarily to be equated with its sound), this must remain an open question. See Ronconi 1963,19-25. notions double

CRAFTED SPEECH IN DE RERUM NATURA

579

in smells, tastes, sounds, and in the experience of pleasure particularly and pain (2.442-43). In fact, variation in sound or smell or taste is largely restricted to an opposition between and pain, expressed as pleasure smoothness or roughness.109 While the quality of the primordia vocum disappears at the level of from the studies of Deutsch, and Friedlander, speech, what emerges is Lucretius' attention to the sensual qualities of the word in the Snyder mouth or in the ear before or, rather, while it gives way to what "under? lies" it.110 These two levels of perception are emphasized by Lucretius himself ("si tibi forte animum tali ratione tenere / versibus in nostris possem, dum perspicis ...," 1.948-49 = 4.23-24).The resulting synaestheof the word's or the line's effects into a apprehended suggest. by logos alone, as the poetics of Philodemus it is a stereoscopic effect achieved the specific Rather, by exploiting nature of sound to register differences in the quality of experience, that does not necessarily map onto atomic reality in the same experience but which, for all that, may not be way as a more "mimetic" knowledge without meaning, alongside discursive claims.111 Regardless of how we sense account for what Lucretius thought he was doing with this manipulation of the phonic quality of his verses, a topic beyond the scope of this it seems wise, in closing, not to concede so readily the subordipaper,112 nation of the phonic materiality of the word to its power to provoke 109 while it is, technically speaking, impossible for the shapes of the atoms them? to be felt, there is a consistent correlation between particular shapes and particular selves sensation (see 2.398-477; 4.542-48,615-32,673-86; cf. 6.773-76, where differences in sensa? tion come about "propter dissimilem naturam dissimilisque / texturas inter sese primasque figuras"). Lucretius blurs in maddening ways the line between primordia and atomic clusters, consistently acting as though the phenomena he adduces to support his arguments displayed features of the former rather than the latter. Perhaps we should understand a relationship like that between the motes in the sunbeam and the atoms (2.138-41), where analogy is also a causal relation in which a similar behavior or form is repeated at each level. This relationship between the shapes of the primordia and the quality of smell or taste or sound is complicated by the need to take into account the differently sized pores of different percipients (4.649-51). See Furley 1993; Graver 1990. 110 See among many possible examples, 2.398-99: "huc accedit uti mellis lactisque liquores / iucundo sensu linguae tractentur in ore."The lines occur in an extended explana? tion of how the shapes of the atoms determine the qualities we experience, and, as Friedlander notes, "the elements of the words appeal to the tongue and the ears as the atoms of the corresponding things appeal to the tongue and taste" (1941, 23). 111 Lucretius' On synaesthesia in metaphors, see Clay 1996, 784-85. 112 is difficult to endorse Friedlander's It position that similarities of sound reflect a greater affinity at the level of ontology. The Stoics are the etymologists, believing that the word could capture something true about the object. Epicureans, as we have seen, appear sia is not so much the unification

580

BROOKE HOLMES between elements and sensation at the word that is heard, and between of the secondary simulacrum, what

images and thought. For there is play the level of the primary simulacrum, elements and imagination at the level the word calls to mind. And herein

lies the double inheritance of lan? with its difference in the sound of its soundings and the things it guage, In marking a certain intransigence communicates. in the suaviloquens carmen that sustains a sense of synaesthesia rather than a sense of that is, in marking the taste of honey that cannot pass its coherence, we may conclude that the word's pleasure beyond the mouth (4.626-29), natural failure of iconic resemblance to the object has value. For this that the object remains a cause of pathe and intransigence guarantees not simply an object of contemplation.113 Or we may choose, with to respect the pleasure of surfaces.114 Lucretius, simply University of North Carolina e-mail: baholmes@email.unc.edu

more interested in dispensing with the nomothetes than in establishing utterances as clues to the nature of things. Thus Origen links the followers of Epicurus with those of Aristotle as people who consider naming an irregular practice (pragma asustaton, Cels. 1.24). See Dalzell 1987,26-28, who makes this argument against Friedlander. Nevertheless, Lucretius may very well be showing off his facility with etymology as a Roman poet, for whom it represents an independent tradition (West, 1969, 97). 113 Asmis 1984,84-98, on See pathe. She understands them as giving us an awareness of inner conditions, as opposed to aistheseis,which give us an awareness of things external to ourselves (97-98); see also Glidden 1979. This seems essentially correct, although it is impor? tant, I think, to stress the role oi pathe in registering what happens at the interface between the body and the world, as in taste and touch but also hearing and smell, which focus on the quality of the effluences rather than on the morphology of the object. The attempt to adapt the vocabulary of atomic shape and pores to vision produces the strange lines 4.706-21, where the only effect of the shape of the primordia is to prevent vision and cause pain by staying in the eyes; the simulacra that cause vision pass through "freely" and never register their presence, only that of the object (718-21; see Bailey 1947 ad 706-21,111.1263-64; Graver 1990). Pathe are largely seen as a criterion for action (the pursuit of pleasure and avoidance of pain) rather than a criterion of truth, but Asmis argues that Epicurus also believed that pathe, like aistheseis, served as a basis for inference about the world. 114 Many thanks to David Armstrong, Jim Porter, Denis Feeney, Christian Wildberg, Jerry Passannante, and the anonymous reader at AIP for their valuable suggestions, interventions, and feedback on this article and its earlier incarnations. Special acknowledgement is due to Katharina Volk, who has been generous with her time, insight, and support from the beginning, and to Jake Mackey for allowing me to use his work on Philodemus and for stimulating discussion. An earlier version was presented at the "Art and Artifice in the Roman World" conference at Harvard University in March, 2002; I thank the organizers for their hospitality as well as the other participants for their comments. I am responsible for remaining errors.

CRAFTED SPEECH IN DE RERUM NATURA BIBLIOGRAPHY

581

Armstrong, David. 1995. "The Impossibility of Metathesis: Philodemus and Lucretius on Form and Content in Poetry." In Obbink 1995, 210-32. Armstrong, David, Jeffrey Fish, and J. I. Porter. Philodemus, On Poems 5, critical ed. and trans. with facing text, notes, and introductory material. Oxford: Clarendon Press. Forthcoming. Arrighetti, Graziano. 1972. Epicuro. Opere. Turin: Giulio Einaudi. Asmis, Elizabeth. 1984. Epicurus' Scientific Method. Ithaca, N.Y.: Cornell Univer? sity Press. -. 1995a. "Epicurean Poetics." In Obbink 1995,15-34. -. 1995b. "Philodemus on Censorship, Moral Utility, and Formalism in Poetry." In Obbink 1995,148-77. -. 1996. "Lucretius on the Growth of Ideas." In Giannantoni and Gigante 1996,11.763-78. Bailey, Cyril, ed. 1922. Lucreti De rerum natura libri sex. Oxford Classical Texts. 2d ed. Oxford: Oxford University Press. ed. and trans. 1926. Epicurus, the Extant Remains. Oxford: Clarendon -, Press. -. 1928. The Greek Atomists and Epicurus. Oxford: Clarendon Press. -. 1947. Titi Lucreti Cari De rerum natura libri six. 3 vols. Oxford: Clarendon Press. Barnes, Jonathan. 1996. "Epicurus: Meaning and Thinking." In Giannantoni and Gigante, 1.197-220. Biville, Frederique. 1997. "La voix signifiante." In Les structures de Toralite en latin (Lingua Latina 4), ed. J. Dangel and C. Moussey, 147-54. Paris: Presses de l'Universite de Paris-Sorbonne. des organes pho-. 2001. "La production de la voix: Anatomie-physiologie nateurs dans les textes latins." In (<Docente natura": Melanges de medecine ancienne et medievale offerts a Guy Sabbah, ed. A. Debru and N. Palmieri, 15-42. St. Etienne: Publications de l'universite de St. Etienne. Blank, David. 1995. "Philodemus on the Technicity of Rhetoric." In Obbink 1995, 178-88. Blickman, D. R. 1989. "Lucretius, Epicurus and Prehistory." HSCP 92:157-91. Boyance, P. 1947. "Lucrece et la poesie." REA 49:88-102. Brunschwig, Jacques. 1977. "Epicure et le probleme du 'langage prive."' Revue des sciences humains 166:157-77. Rpt. "Epicurus and the Problem of Pri? vate Language." In Papers in Hellenistic Philosophy. Trans. Janet Lloyd, 2138. Cambridge: Cambridge University Press. Buchheit, V. 1984. "Lucrez uber den Ursprung von Musik und Dichtung." RhM. 127:141-58. Campbell, Gordon. 2003. Lucretius on Creation and Evolution. A Commentary on De Rerum Natura, Book Five, Lines 772-1104. Oxford: Oxford Univer? sity Press.

582

BROOKE HOLMES

Chilton, C. W. 1962. "The Epicurean Theory of the Origin of Language." AIP 83:159-67. Classen, C. J. 1968. "Poetry and Rhetoric in Lucretius." TAPA 99:77-118. Clay, Diskin. 1996. "An Anatomy of Lucretian Metaphor." In Giannantoni and Gigante 1996,11.779-93. Cole, Thomas. 1967. Democritus and the Sources of Greek Anthropology. Philological Monograph XXV of the American Philological Association. Cleveland: Scholars' Press. Dalzell, Alexander. 1987. "Language and Atomic Theory in Lucretius." Hermathena 143:19-28. de Lacy, Phillip. 1939. "The Epicurean Analysis of Language." AJP 60:85-92. Deleuze, Gilles. 1990. "Lucretius and Naturalism." In The Logic of Sense. Trans. Mark Lester, 266-79. New York: Columbia University Press. Deutsch, R. E. 1978. The Pattern of Sound in Lucretius. New York: Garland. Diels, Hermann. 1899. Elementum. Leipzig: Teubner. Dionigi, Ivano. 1988. Lucrezio: Le parole e le cose. Bologna: Patron. Everson, Stephen. 1994a. "Epicurus on Mind and Language." In Everson 1994b, 74-108. ed. 1994b. Language: Companions to Ancient Thought 3. Cambridge: -, Cambridge University Press. Ferguson, J. 1987. "Epicurean Language-Theory and Lucretian Practice." LCM 12:100-105. Frede, Michael. 1994. "The Stoic Notion of a lekton." In Everson 1994b, 109-28. Friedlander, Paul. 1941. "Pattern of Sound and Atomistic Theory in Lucretius." AJP 62:16-34. Furley, David. 1971. "Knowledge of Atoms and Void in Epicureanism." In Essays in Ancient Greek Philosophy, ed. J. P. Anton and G. L. Kustas, 1.607-19. Albany: State University of New York Press. -. 1993. "Democritus and Epicurus on Sensible Qualities." In Passions & Perceptions. Studies in Hellenistic Philosophy of Mind. Proceedings of the Fifth Symposium Hellenisticum, ed. Jacques Brunschwig and Martha Nussbaum, 72-94. Cambridge: Cambridge University Press. Gale, Monica. 1994. Myth and Poetry in Lucretius. Cambridge: Cambridge Uni? versity Press. Giannantoni, Gabriele, and Marcello Gigante, eds. 1996. Epicureismo greco e romano: Atti del congresso internazionale, Napoli, 19-26 maggio 1993. 2 vols. Naples: Bibliopolis. Giussani, Carlo, ed. 1896-98. T. Lucreti Cari De rerum natura libri sex.Txxx'm:Casa Editrice G. Chiantore. Glidden, D. K. 1979. "Sensus and Sense Perception in the De Rerum Natura." CSCA 12:155-81. -. 1983. "Epicurean Semantics." In EYZHTHIII: Studi sulTEpicureismo Greco e Latino offerti a Marcello Gigante, 185-226. Naples: Gaetano Macchiaroli Editore.

CRAFTED SPEECH IN DE RERUM NATURA -.

583

1985. "Epicureanprolepsis." Oxford Studies in Ancient Philosophy 2:175215. Godwin, John, ed. and trans. 1986. Lucretius: De rerum natura IV. Wiltshire: Aris & Phillips. Goldschmidt, Victor. 1978. "Remarques sur l'origine epicurienne de la 'prenotation.'" In Les Sto'iciens et leur logique: Actes du colloque de Chantilly, 1821 septembre 1976.155-69. Paris: J. Vrin. Gourevitch, Danielle. 1997. "La voix et la parole, Lucrece, Galien et... Colette." Revue Officielle de la Societe Frangaise d'O.R.L. et de pathologie cervicofaciale 45:15-18. Graver, Margaret. 1990. "The Eye of the Beholder: Perceptual Relativity in Lucretius." Apeiron 23:91-116. di Hammerstaedt, Jiirgen. 1996. "II ruolo della Epicurea nell'interpretazione nPOAH^IZ Epicuro, Epistula ad Herodotum 37sg." In Giannantoni and

Gigante 1996,1.221-37. Hinds, Stephen. 1987. "Language at the Breaking Point: Lucretius 1.452." CQ 37:450-53. Holmes, Brooke. "Narcissus in Mourning: Three Texts on the Attractions and the Limits of Self-Mirroring (Aeschylus, Lucretius, Ovid)." In Finestre. Quaderni di Synapsis V. Atti della Scuola Europea di Studi Comparati. Ed. Letizia Bellocchio. Florence: Le Monnier. Forthcoming. Jensen, Christian, ed. 1923. Philodemos uber die Gedichte, Filnftes Buch. Berlin: Weidmann. Kemke, Johannes, ed. 1884. Philodemi de musica librorum quae exstant. Leipzig: Teubner. Kennedy, Duncan F. 2002. Rethinking Reality: Lucretius and the Textualization of Nature. Ann Arbor: University of Michigan Press. Kleve, Knut. 1989. "Lucretius in Herculaneum." CErc. 19:5-27. Koenen, M. H. 1997. "Lucretius' Olfactory Theory in De rerum natura IV." In Lucretius and his Intellectual Background, ed. K. A. Algra, et al., 163-77. Amsterdam and New York: North Holland. -. 1999a. "Lucretius' Explanation of Hearing." Mnemosyne ser. 4.52:434-63. -. 1999b. "Some Aspects of Lucretius' Account of Hearing." In Presence de Lucrece: actes du colloque tenu a Tours, 3-5 decembre 1998, ed. Remy Poignault, 21-36. Tours: Universite de Tours. Konstan, David. 1973. Some Aspects of Epicurean Psychology. Leiden: Brill. Lee, E. N. 1978. "The Sense of an Object: Epicurus on Seeing and Hearing." In Studies in Perception, ed. P. K. Machamer and R. G. Turnbull, 27-59. Columbus: Ohio State University Press. Lloyd, G. E. R. 1966. Polarity and Analogy. Cambridge: Cambridge University Press. Long, A. A. 1971. "Aisthesis, Prolepsis and Linguistic Theory in Epicurus." BICS 18:114-33.

584

BROOKE HOLMES

Long, A. A., and D. N. Sedley. 1987. The Hellenistic Philosophers. Vols. 1 and 2. Cambridge: Cambridge University Press. Mackey, Jacob. 2003. "New Evidence for the Epicurean Theory on the Origin of Language." M.St. diss., Christ Church, Oxford. Mangoni, Cecilia. 1992. "II PHerc. 407 della Poetica di Filodemo." CErc. 22:13137. ed. and trans. 1993. // Quinto Libro della Poetica (PHerc. 1425 e 1538). -, Naples: Bibliopolis. Monet, Annick. 1996. "Philodeme. Sur les sensations. PHerc 19/698." CErc. 26:27126. Morris, Sarah. 1992. Daidalos and the Origins of Greek Art. Princeton: Princeton University Press. Munro, H. A. J., ed. 1893-1900. T Lucreti Cari De rerum natura libri sex. 4th ed. Cambridge: Deighton Bell. Obbink, Dirk, ed. 1995. Philodemus and Poetry: Poetic Theory and Practice in Lucretius, Philodemus, and Horace. Oxford: Oxford University Press. -. 1996. "Epicurus on the Origin of Poetry in Human History." In Gian? nantoni and Gigante 1996,11.683-700. Offermann, Helmut. 1972. "Lukrez V 1028-1090." RhM 114:150-56. Pigeaud, J. M. 1980. "La physiologie de Lucrece." REL 58:176-200. -. 1983. "Epicure et Lucrece et l'origine du langage." REL 61:122-44. Porter, J. 1.1989. "Philodemus on Material Difference." CErc 19:149-78. -. 1996. "In Search of an Epicurean Aesthetics." In Giannantoni and Gigante 1996,11.611-28. . 2003. "Lucretius and the Poetics of Void." In Lejardin romain: Epicurisme et poesie a Rome. Melanges offerts a Mayotte Bollack, ed. Annick Monet, 197-226. Lille: Universite Charles-de-Gaulle Lille 3. Puglia, Enzo. 1988. Aporie testuali ed esegetiche in Epicuro (PHerc. 1012) / Demetrio Lacone. Naples: Bibliopolis. Ronconi, Alessandro. 1963. "Appunti di estetica epicurea." In Miscellanea di studi Alessandrini in memoria di Augusto Rostagni, 7-25.Turin: Bottega d'Erasmo. Rosenmeyer, Thomas. 1996. "Sensation and Taste in Lucretius." SCI 15:135-51. Sbordone, Francesco. 1971. "Ancora un papiro ercolanese della Poetica di Filodemo: N. 403." In Studi filologici e storici in onore di Vittorio de Falco, 343-52. Naples: Libreria Scientifica Editrice. Schiesaro, Alessandro. 1990. Simulacrum et imago: Gli argomenti analogici nel De rerum natura. Pisa: Giardini. -. 1994. "The Palingenesis of De Rerum Natura." PCPS 40:81-107. Schrijvers, P. H. 1970. Horror ac divina voluptas. Etudes sur la poetique et la poesie de Lucrece. Amsterdam: A. M. Hakkert. -. 1974. "La pensee de Lucrece sur l'origine du langage (DRN V.10191090)." Mnemosyne ser. 4:27:337-64. Rpt. in R H. Schrijvers. 1998. Lucrece et les sciences de la vie. Mnemosyne, bibliotheca classica Batava. Supplementum, 186. Boston: Brill.

CRAFTED SPEECH IN DE RERUM NATURA

585

Sedley, David. 1973. "Epicurus. On Nature, Book XXVIII." CErc 3:5-83. -. 1989. "Epicurus on the Common Sensibles." In The Criterion of Truth. Essays in Honor of George Kerferd, ed. Pamela Huby and Gordon Neal, 123-36. Liverpool: Liverpool University Press. -. 1998. Lucretius and the Transformation of Greek Wisdom. Cambridge: Cambridge University Press. Serres, Michel. 2000. The Birth ofPhysics. Trans. Jack Hawkes. Manchester: Clinamen Press. Smith, M. F. 1993. Diogenes of Oinoanda: The Epicurean Inscription. Naples: Bibliopolis. Snyder, J. M. 1980. Puns and Poetry in Lucretius' De Rerum Natura. Amsterdam: Gruner. Sudhaus, Siegfried, ed. 1892-96. Philodemi Volumina Rhetorica. 2 vols. Leipzig: Teubner. Thury, E. M. 1987. "Lucretius' Poem as a Simulacrum of the Rerum Natura." AJP 108:270-94. Usener, Hermann, ed. 1887. Epicurea. Leipzig: Teubner. Volk, Katharina. 2002. The Poetics of Latin Didactic. Oxford: Oxford University Press. Vlastos, Gregory. 1946. "On the Pre-History in Diodorus." AJP 67:51-59. West, David. 1969. The Imagery and Poetry of Lucretius. Edinburgh: Edinburgh University Press. West, M. L. 1965. "Tryphon De Tropis." CQ 15:230-48. Wigodsky, Michael 1995. "The Alleged Impossibility of Philosophical Poetry." In Obbink 1995, 58-68.