James Elkins On the Impossibility of Close Reading
    
    I take it that an interest in the detail is not just a specialty of some art history, and not just an advanced subject suitable for monographic studies, but a typical trait of twentieth- and twenty-first century art history. I take it some of the most challenging and innovative art history is written in thrall of the detail: we are not only curious about it, we invest it with the capacity to unlock entire artistic practices, to reveal the way beyond critical impasses, to reach deeper strata of pictorial meaning, to show how images themselves are constituted in their historical, psychological, and formal origins. I also take it that the detail is a characteristically modern interest: that much can be proven by comparison with texts written before the midtwentieth century, which seldom mention details within artworks.1 Given those investments, and given that we live inside the era of the detail, how much of our interest can we understand? My contribution here is excerpted from an inquiry into close reading — not quite the theme of the detail, but also inextricable from it — that I wrote as part of an investigation of art historical interpretation. My example in the essay is prehistoric marking and my starting point is Carlo Ginzburg’s famous essay. Those are, to some degree, arbitrary starting points. I will draw some conclusions from those choices at the end. But first I would like to take a page or two and try to understand the phenomenon of the fascination with the detail more broadly. Let me propose five reasons we remain hypnotized by the detail. 1. The detail can be understood as a rupture in the symbolic order of art history. Since the spread of iconographic analysis inspired by Panofsky model, art historical interpretation has had an investment in producing symbolic orderliness. An iconographic interpretation
    1
    
    This is argued at length in my Why are Our Pictures Puzzles? On the Modern Origins of Pictorial Complexity (New York: Routledge, 1999).
    
    involves locating, labeling, and classifying an artwork’s range of symbolic references, and an ideal iconographic interpretation is complete: it is both systematic and thorough. It brings order to the work’s symbolic meanings, and it creates a sense that symbols are themselves potentially orderly. Even though iconography is not a method that many art historians would be likely to endorse (most of us would probably rather say we do not practice any particular method, or that we owe allegiance to something else, such as semiotics or poststructuralism), I think it can be argued that iconography remains the most common method, in practice, throughout the world. A detail can be said to rupture the symbolic order of art history by breaking iconography’s grip on art historical method (it is a cold grip, and very strong!), piercing the symbolic orderliness of ordinary interpretation and promising something beyond it. In this respect, a fascination with the detail can be nothing more or less than an attempt, sometimes not fully articulated, to escape the potentially rigid grip of iconographic interpretation. On the other hand, details can help find different structures of meaning in artworks. In this capacious, promising, and ill-defined territory, I would name Hubert Damisch (the /cloud/ is one of these) and, very differently, the work of the honoree of this book, Friedrich Teja Bach. At the least, odd “structural detail” like the emblematic / naturalistic / calligraphic swirls in Dürer’s prints, or certain repeated structures in Cézanne, are candidates for a different kind of looking, one that would divide the images differently.2 (In the spirit of this Festschrift let me add, parenthetically, how refreshing I find it to read Teja Bach’s texts: his
    2
    
    Friedrich Teja Bach, “Albrecht Dürer: Figures of the Marginal,” Res 36 (1999): 79-99; Bach, “The Stake in the
    
    Pictorial Flesh: Disruptions in Cézanne’s Oeuvre,” in Cézanne: Finished / Unfinished, exh. cat., edited by Felix Baumann et al. (Kunstforum, Vienna; Kunsthaus, Zürich: Hatje Cantz / Distributed by D.A.P., 2000), 63-83. There is much more to be done with Cézanne along the lines Bach investigates; I have written on related ideas in “The Failed and the Inadvertent: The Theory of the Unconscious in the History of Art” International Journal of Psycho–Analysis 75 part 1 (1994): 119–32; a revised version (2001) is online at
    
    www.academyanalyticarts.org/elkins.htm.
    
    analytical sharpness is undulled by the rote allegiance so much Anglo-American scholarship feels it needs to pay to identity politics, gender, and social meanings. Only with those relentlessly generalizing, and often non-visual, impediments out of the way, is it possible to take the time to see as clearly as Teja Bach does. Happy sixtieth!) 2. To return to my informal list: the detail can also be seen as evidence of a primal visual illegibility. Throughout modernism there has been a temptation to picture visual art as a region uninhabited by narrative, “code,” semiotics, language, or logic. The “word-image dichotomy” that seemed so important in the 1970s and 1980s (and which lives on in the International Society for Word and Image) has been vigorously but perhaps ineffectively opposed by poststructuralist critiques such as Mieke Bal’s book on Rembrandt, subtitled Beyond the Word-Image Opposition. Such critiques are ineffective because the reified, pure realm of the “merely” visual is a structural component of modernism. As long as modernist ideas continue, the visual will be a place that is seen, at least by some, as a realm apart from language, where experience is uncognized, nonconceptual, subsemiotic, or otherwise distilled and abstracted from articulable meaning. I have fallen into that desire myself, and I have been taken to task for it by the Renaissance specialist Robert Williams.3 In a book called On Pictures, And the Words That Fail Them, I tried to find theories that would go to the brink of the nonconceptual. Williams pointed out that hope placed in the domain of the visual can be historicized as an artifact of a certain sense of modernism, and that it is therefore undependable and unhelpful for art historical analyses.4 Williams is partly right and partly wrong: he is correct to want art historians to see their places in history more clearly, and to stop hoping for a transhistorical or extra-longuistic realm of pure color and form. He is probably wrong, however, to imply that
    
    3
    
    [Williams’s review] Williams, [ ] I do not want to imply that are currently collaborating on a book on Renaissance historiography.
    
    4
    
    See also his essay on my work, [ ]
    
    we can think our way out of our desire for such an escape, or talk ourselves out of a lingering belief in the immaculately obscure domain of the visual. The detail remains the privileged example of pure visuality, because it is what apparently remains when the “codes” (as Roland Barthes said) are subtracted away from an image. If, following a logic that recurs throughout modernism, we mentally subtract whatever can be assigned a meaning, we are left with “pure” color, “pure” motion, “pure” form — lines, shapes, impressions — that then appear to be nothing other than the transcendental ground of visuality itself. It is a powerful dream, and it is not one we are likely to wake up from anytime soon. 3. The detail can be seen as the object of a properly scientific investigation. Just as a scientist looks into an atom, or a doctor into an ear, or a biologist into a microscope, so an art historian can peer into a picture. In this third model, when art historians concentrate on details they are only doing what scientists are charged with doing: they are systematically dissecting or disassembling their objects into component parts. Just as a mycologist looks at spores, or an ornithologist at the markings on birds’ breasts, or a dermatologist at tiny suspicious spots, so an art historian looks at details. This third model is art history’s version of the common desire to become scientific, a desire that has long infected the humanities. There is probably no perfect cure for this malaise because art history is predicated, to some degree, on the idea that it is not scientific, and that denial prevents us from talking openly about our scientific hopes or pretensions. I have little to say about this third model, because I think I suffer from it about as much as anyone. It is pandemic in the humanities, but like cancer it is not something we often like to talk about. Speaking for myself, I certainly try not to sound scientific when I deploy details in a scientific manner. 4. The detail can be valued in accord with the Romantic aesthetic of the fragment. As Guy Sircello, Barbara Stafford, and others have argued, the fragment played a central role in
    
    Romantic aesthetics; it was taken to possess a greater immediacy than the whole, as well as a privileged relation to truth.5 Artworks were understood to have been muted by systems of academic conventions and skills, and by concepts such as balance, symmetry, composition, and especially decorum. Details were thought to be outside such systems. Just as my second model is evidence of the continued influence of modernism, so an interest in the fragment demonstrates the part that Romanticism still plays in contemporary understanding of artworks. The sign that an interpretation of a detail is driven by Romantic ideas is the notion that the detail is closer to the artist’s intention or subjectivity, or more reliable as a vehicle of expression, than the whole of which it is a part. An anti-Romantic position exists in art history, but I think that remnants of Romanticism are much more common than anti-Romantic reactions. Williams has recently suggested that Renaissance specialists consider what he calls “the systematicity of art,” its structural dependence on an overall internal and semantic order.6 Against that initiative there is, however, a large literature that finds details to be more reliable indices to an artwork’s meaning. 5. And finally, the detail is also an object of interest in surrealism and psychoanalysis. The interest in the pathological and the non-normative is a characteristic of surrealism, and it is entwined with the emphasis placed by psychoanalysis on the symptom. This theme is ramified and not easy to characterize as a whole; it includes a generalized interest in the pathological, but also the surrealist “part-object,” Jacques Lacan’s paradoxical objet petit a, and Julia Kristeva’s desired and discarded abject.7 The histories of psychoanalysis and surrealism are entwined and have a broad conceptual and temporal range: both surrealism and
    
    5
    
    See my review of Barbara Stafford, Body Criticism (MIT, 1991), The Art Bulletin 74 no. 3 (1992): 517–520. Williams, [ ] For the part-object see Part Object Part Sculpture, exh. cat., edited by Helen Molesworth, with essays by
    
    6
    
    7
    
    Rosalind Krauss, Briony Fer, et al. (University Park, PA: Penn State Press / Wexner Center for the Arts, 2005).
    
    psychoanalysis are very much still alive in the art world, as well as in the interpretive agendas of contemporary historians and theorists.8 In this context I want to address just one writer, Georges Didi-Huberman. His work has connections with each of the five categories I am proposing, but I think his deepest affinities lie here. In l’Image survivante, Didi-Huberman is concerned with Aby Warburg’s interest in the possibility of a “pathology” of art rather than the normative humanist discipline that art history continues to be.9 Didi-Huberman has long been interested in moments in which the object of art historical attention becomes illegible, and in l’Image survivante, art history’s preternatural certainty about its objects is theorized as a repression of the eruptions of the irrational that had concerned Warburg. In this reading, if a figure or form is of compelling interest — if it sidetracks an interpretation, or emerges unexpectedly in an otherwise orderly exposition — then it may be understood in the way Warburg understood Pathosformeln, recurrent, inadequately repressed figures that emerge and re-emerge in normative historical accounts. Such disturbing, cathected figures are a destabilizing force within art history, with the potential to recall for us the labile meanings of images, and to reconfigure what counts as historical sequences and symbolic orders. Didi-Huberman’s analysis depends on psychoanalysis, and his affective interests spring from surrealism. Devant l’image, and especially its final chapter, in English as “Appendix: The Detail and the Pan,” are replete with surrealist affect: the pan is allied with blindness, bedazzlement, difficulty, with the uncontrolled, the accidental, with madness, A good measure of this is Hal Foster’s ambiguous endorsement of psychoanalysis as both a tool and an historical fact in his unsigned introduction to Foster, Yve-Alain Bois, Rosalind Krauss, and Benjamin Buchloh, Art Since 1900: Modernism, Antimodernism, Postmodernism (London: Thames and Hudson, 2004); see also my review of Foster, Compulsive Beauty (MIT, 1993), The Art Bulletin 76 no. 3 (1994): 546–48.
    8
    9
    
    I have other remarks on Didi-Huberman’s larger project in an essay on indeterminacy, forthcoming in
    
    Bestimmte Unbestimmtheit: Perspektiven philosophischer Ästhetik, edited by Gerhard Gamm und Eva Schürmann (2006).
    
    stains, the erratic, with unexpected intrusions, with radiance, with surging, crisis, with the non-mimetic, with imprecision, the “quasi” and the marginal, the unsettling, the uncanny, the infectious.10 I offer this not as a summary of Didi-Huberman’s argument regarding the pan and the detail, but to conjure the way his theory flows (another word that is important in his critique) from surrealist tropes. His revision of “a certain art history” intersects all five of my heuristic categories, but it takes its energy from desires in regard to the image that can be traced, in specific lineages, to surrealism and psychoanalysis. I do not mean this in any way as a criticism: he is an acquaintance, and I have found many parallels between our projects over the last ten years. It may be that he is the most forceful embodiment of an alternative to high modernism that is well known in art, but whose consequences in historiography are only now being discovered. What I have to say in this essay is substantially different from this fifth category, although it is also entwined in all five. It is a case of hypertrophied interpretation. Its pathology is mainly known to its participants (including, in part, its main practitioner, Alexander Marshack), and the discussion that followed the original publication of the essay, which involved published letters from a dozen art historians and archaeologists, did not result in any reconfiguration of the discipline.11 Let me say in the spirit of friendship that this essay may seem insufficient or superficial in Didi-Huberman’s terms, more like a persistent rash on
    
    10
    
    Didi-Huberman, Confronting Images: Questioning the Ends of a Certain History of Art, translated by John
    
    Goodman (University Park, PA: Pennsylvania State University Press, 2005). All these words are from pp. 25256, in the discussion of Vermeer’s The Lacemaker. 11 The following essay is abridged, by Wolfram Pichler, from the chapter “On the Impossibility of Close Reading,” originally published as “On the Impossibility of Close Reading: The Case of Alexander Marshack,” Current Anthropology 37 no. 2 (1996): 185–226, and then, in an expanded form, as a chapter in my book Our Beautiful, Dry, and Distant Texts: Art History as Writing (University Park, PA: Penn State Press, 1997); paperback edition, with new preface (New York: Routledge, 2000). The version in Anthropology Today includes responses to the essay by a number of historians, including T.J. Clark and Whitney Davis.
    
    the skin of art history than a traumatic memory from the unconscious of the discipline. But there is room in a field as capaciously blind as art history for several rival diagnoses. What follows is an abridged version of the essay. At the end I will say which of the five models I think apply best to the case at hand.
    
    1 Close reading of visual images is a constant ideal in art history and criticism—it is virtually never questioned, and in general it seems to be a good idea in any field, for any purpose, and under any theoretical regimen. Its genealogy begins where art history does, with antiquarianism and connoisseurship in the seventeenth and eighteenth centuries, and ascends through nineteenth–century Morellian methods, branching and proliferating in the twentieth century into style analysis, formalism, and iconography.12 Carlo Ginzburg’s over–read essay on Morelli, Freud, and Sherlock Holmes remains the principal source for the history of the idea, though Ginzburg’s concatenation of connoisseurship, psychoanalysis, art history, and detective novels could have been expanded by consideration of the way the same term is used in philology and literary criticism—including Ginzburg’s own essay, which becomes an example of the method it exposits.13 It is questionable how much sense it makes to call such far–flung practices a “method,” as Ginzburg does; instead I would say that different methods grow from the enabling concept of close reading. The transparent nature of that concept would seem to forestall any direct analysis or critique—whatever a close reading may be in practice, the concept of close reading remains nearly intangible. Accuracy and even insight are intuitively tied to the closeness of a reading, regardless of disputes about evidence or theory. The
    Among other sources, see Roland Kany, “Lo sguardo filologico, Aby Warburg e i Dettagli,” Annali della Scuola Normale Superiore di Pisa, Classe di Lettere e Filosofia series 3, vol. 15, no. 3 (1985): 1265–1283. 13 Carlo Ginzburg, “Morelli, Freud, and Scherlock Holmes: Clues and Scientific Method,” History Workshop 9 (1980): 1 ff. Ginzburg is fully aware of his attraction to the method, but only partly aware of the reasons why he is so attracted. His books, I think, replay the results of that attraction by applying the method to various subjects, rather than delving into the question of the place of his writing within the tradition, and the limits of self– reflexivity.
    12
    
    unobjectionable universality of close reading is most apparent in our inability to think what kind of visual artifact (or text, or crime scene, or narrative) might not be best understood by close examination. In short the purposive, vigilant scrutiny of every mark or sign seems to be a foundational principle of understanding, whether the object is the most recent installation art or the oldest markings on stone. After close reading, reading can relax, and there are many ways to interpret quickly, to make epitomes and outlines, to paraphrase and abbreviate, to schematize and summarize and condense and abstract: but each one of them is dependent on the possibility of a logically prior close reading. When readings are questioned for their closeness or sloppiness, the purpose is usually to elicit the historians’ covert agendas, and to show how theory determines what and how we see. But that kind of critique has more to do with the construction of history than with close reading itself. It is possible to doubt that someone’s reading is as close as it might be, or to claim it is skewed by prejudiced seeing, and at the same time never begin to doubt close reading itself. Here I will be trying to open a critique of close reading, rather than the purposes to which it is put. In the end I will argue that in effect close reading does not exist: in the strict sense it is impossible because any reading is comprised of echoes of looser accounts and suppressed promises of even closer readings. I mean that no close reading has ever been made, because what we understand as a close reading is a promise, something that is evoked and intimated, a specter that can only be grasped by conjuring and then repressing other kinds of reading. That which appears as close reading in innumerable texts—including Ginzburg’s —is a moment of incomplete awareness, built on self–contradiction and the resurgent hope of intimacy with the object. As an example I take the work of the archaeologist and art historian Alexander Marshack, and I would like to imply—though I will not be arguing this directly—that Marshack’s methods make him one of the most important art historians of the century, one whose work deserves the attention not only of all art historians, but also of anyone involved in
    
    the project of seeing and interpreting as finely as possible. 14
    
    Marshack’s questions are
    
    especially exemplary for art historians who are engaged in analyzing gestural brushstrokes, invidual marks, color areas, facture, and other phenomena that are taken to be minimal components of pictorial meaning—lexemes, in the semiotic term, out of which the larger signifying units of pictures are built. Semiotic art history in particular needs to be clear about the place where “subsemiotic” signs leave off and minimally semiotic signs begin; and I will suggest at the end that Marshack’s analyses press that point as no others have done.15 Marshack’s approach takes some introducing; after examining one of his analyses, I will outline the kinds of response he has received.
    
    2 Mashack’s accounts of marks on Upper Paleolithic and Mesolithic bones, slate, and amber are among the most careful analyses in all of archaeology as well as in art history and criticism, visual theory, connoisseurship, and conservation.16 He looks closely, literally with a microscope, but unlike painting conservators he also looks at every mark on a surface or artifact, and his looking does not cease until he has satisfied himself that he has distinguished all intentional marks from unintentional or random marks, ordered the intentional marks in chronological sequence, distinguished directions in which marks were made, noted where tools were lifted from the surface and where they remained in contact, and determined how many tools or cutting edges were used to make the marks. His analyses are lessons in looking: forcible patient attempts to see everything, together with a concerted effort to conclude nothing that cannot be empirically demonstrated.
    
    See “Before Theory,” my review of Whitney Davis, Masking the Blow, inArt History, forthcoming, 1993. For a way of distinguishing subsemiotic from semiotic and supersemitic (i.e., aggregative) signs, see Mieke Bal, Reading “Rembrandt,” Beyond the Word–Image Opposition (Cambridge: Cambridge Universiy Press, 1991). 16 Alexander Marshak, The Roots of Civilization: The Cognitive Beginnings of Man’s First Art, Symbol and Notation (New York: McGraw–Hill, 1972). The second edition (Mount Kisco, N.Y.: Moyer Bell Ltd., 1991) contains several significant revisions and a new section at the end.
    14 15
    
    Marshack’s central claim is that many Paleolithic markings—even tiny ones, at the threshold of unaided vision—are notations recording the observational points of the lunar cycle. A notation might be started, for example, at new moon, and continue as an accumulation of marks, grouped according to observational points in the cycle such as half– moons, new moons, or full moons.17 It is a hypothesis that might appear to be readily susceptible to various counterclaims, and in fact Marshack presents it as a reasoned scientific hypothesis of exactly the kind that should be open to empirical falsification. Yet that genuinely scientific attitude has gone largely unread, and scholars have been more struck by the persistence with which he has explicated his thesis over a period of more than thirty years. The archaeologist Denise Schmandt–Besserat, for example, has said that Marshack’s theory of lunar records “cannot be proven or disproven nor can it be ignored.”18 Whitney Davis dedicated a book on late prehistoric Egyptian art to Marshack, whom he says “has been asking questions that I am not sure any archaeology or art history could answer.”19 Donald Preziosi has recently used Marshack’s research uncritically, as an example of the most interesting kinds of questions art history can ask.20 (It is characteristic that these positive assessments are not accompanied by further close readings. Preziosi calls Marshack’s work “emblematic of the poststructuralist critique of verbocentrist structuralism” but he omits a discussion of the “detailed arguments” that would support the point.21 It is a typical effect of
    This is different from the claim, which I had initially taken Marshack to have been making, that the Paleolithic notations record the phases of the moon in the way a modern astronomer might. The point is rather that the Paleolithic engravers had a “tendency” to begin or end their sets of marks at observational points: hence their “completed” artifacts need not be arithmetically accurate, but will tend to dispose themselves in groups according to potentially observable lunar sequences. The lunar model, in turn, is astronomically correct and can therefore be employed to test possible correlations with Paleolithic markings. See Marshack’s reply to “On the Impossibility of Close Reading,” op. cit., p. [12, 13]. 18 D. Schmandt–Besserat, Before Writing, op. cit., 160. 19 Whitney Davis, Masking the Blow: The Scene of Representation in Late Prehistoric Egyptian Art (Berkeley: University of California Press, 1992), xvi. Davis doubts a number of Marshack’s particular methods, assumptions, and conclusions, as he has said in essays in Current Anthropology and elsewhere. But he finds Marshack’s questions often unanswerable within art history as it is constituted. See also my review, “Before Theory,” Art History Art History 16 no. 4 (1993): 647–72. 20 Donald Preziosi, Rethinking Art History: Meditations on a Coy Science (New Haven: Yale University Press, 1989), 133–42. 21 Marshack’s “detailed arguments,” Preziosi writes, “will not be followed here; for our purposes, Marshack’s conclusions and some of their implications will suffice”; and he directs the reader a 1979 exchange in Current Anthropology for an “overview the issues involved.” The problem here is that the millimeter–by–millimeter way
    17
    
    extremely close readings that they pass outside the normal protocols of interpretation and are skipped in more general accounts.) It was praise like this that first attracted me to Marshack’s writing, and what kept my attention was a strange dynamic that has repeated itself over the years as archaeologists have raised objections and Marshack has replied and gone on to make new, even more detailed analyses. The lunar hypothesis often seems patently unlikely, but Marshack’s detractors seem to have no purchase on his methods, and no power to overturn his conclusions. 22 It seems easy to say that Marshack is seeing too much, and trying for example to interpret careless decoration, or meaningless tool–sharpening marks, as deliberated calendrical notation. But when these and other problems are brought up the result is almost invariably the same: the people who object are led to see that they have been operating with certain preconceptions about how accurate marks should be, or what decoration is, or how it is different from notation. What appear to be glaring mistakes in Marshack’s method turn out to be unresolved aspects of our habits of seeing. The critics are silenced, in effect, by unanswerable questions: What should a Paleolithic image look like? What is decoration, and how does it differ from notation? Marshack’s analyses, perhaps more than any others in the history of art, force these questions into the open, and allow us to begin to see how we see, and what we think images are.23
    
    that Marshack reads offers the only possibility of understanding his ostensible methodological breaks with structuralism. See Preziosi, Rethinking Art History, op. cit., 141–42, and 231, nn. 34, 44, citing Marshack, “Upper Paleolithic Symbol Systems of the Russian Plain,” Current Anthropology 20 no. 2 (1979): 271–95, with comments and Marshack’s reply in ibid., 295–309. 22 The work of Francesco d’Errico, for example “Paleolithic Lunar Calendars: A Case of Wishful Thinking?” Current Anthropology 30 (1989): 117, is cited as the most serious challenge to Marshack’s work. d’Errico works with an electron microscope, and some of his objections are cogent; but from the point of view I am adopting here, his analyses raise more problems than they solve. They are extremely idiosyncratic, both graphically and methodologically, and they deserve a separate study rather than a precipitate comparison with Marshack’s work. See also Marshack’s full–scale study, “The Taï Plaque and Calendrical Notation in the Upper Paleolithic,” Cambridge Archaeological Journal 1 no. 1 (1991): 25–61. 23 I am only considering these issues from the point of view of the question of close reading. For recent archaeological and historiographic work on this topic see Whitney Davis, “Beginning the History of Art,” Journal of Aesthetics and Art Criticism 51 no. 3 (1993): 327–50, and the book Replications: Art History, Archaeology, Psychoanalysis (University Park, PA: Penn State Press, 1996).
    
    In what follows I am going to do what has, I think significantly, seldom been done in the literature up to this point: I am going to make a reading of one of Marshack’s readings that is closer than his own reading. Since the original text I will be analyzing is less than a full page (though it is interrupted by illustrations, and occupies parts of four pages), I will be quoting it in full, so it can be read within my text as a sequence of indented paragraphs.24
    
    3 In The Roots of Civilization, from which this example is taken, Marshack sets the stage for each new object with a brief narrative. In this case he has just discussed a large bâton made of a bone, sculpted into a fox’s head at one end, pierced with a hole, and engraved in rows of notches. The text continues:25
    
    In cabinet number one at the Musée des Antiquités Nationales, not far from the large bâton, lies a smaller, fragile, better–worked bit of bone, discolored a dead grey with time. It is the handle of a bâton, with the head and hole broken off, but with the arc of the hole visible at the break. It too comes from Le Placard, apparently from the Magdalenian IV, and may be a few years, a few hundred years, or a few thousand years later than the bâton with the head of a fox. Since it is a fragment, it has not been considered an important find. It is marked with what has been traditionally called decoration in a “geometric” style [Plate 5, top].
    
    The plate numbers in brackets in my quotations replace figure numbers in parentheses in the original text; they are not interpolated references. I have redrawn the line drawings in Marshack’s text. 25 To avoid confusion, since I will be citing the text in its entirety, I have omitted page numbers. The discussion begins on p. 91 of Roots of Civilization and ends on p. 94. Marshack’s photographs have been retained here, though their numbers have been changed in my quotations.
    24
    
    Marshack directs the reader to an overall shot of the front face of the bone (Plate 5, top). Then the reader is invited to move in on the object, and review its markings more attentively (Plate 5, bottom): The bâton has been worked to create four flattened sides, and on one of these there is a raised platform on which the markings were made. The handle itself is unbroken and the markings on it are complete. It is possible, then, that we have all the markings originally made. The next paragraph announces the results of microscopic analysis. Virtually all of Marshack’s studies are done with a microscope; he makes photomacrographs and photomicrographs, and uses them to help modify and complete the notes he makes in the presence of the object. This particular paragraph is interrupted in a spectacular fashion, since the first sentence is interrupted by the end of the page, and on turning the leaf the reader is confronted with a full– page, marginless illustration of a tiny portion of the bone, at a magnification of about 50 times (Plate 7). Marshack continues:
    
    Microscopic analysis of the main face [Plate 5, bottom; Plate 6] revealed sets of parallel lines made by 10 different [engraving] points, with groups that seem to contain “symbols” and “signs.” As a first step in the microscopic analysis Marshack distinguishes several series of marks made by different burins:
    
    At one place a mark at the end of a series of 14 lines, all made by a sharp, thin point, is crossed over by another that point that scratched a flat floor with two
    
    small points at each side (
    
    ) [Plate 6].
    
    The reference here is to the full–page photograph, which is enlarged so much that at first it is difficult to tell that the marks sloping up and to the left are a single stroke of the burin, leaving a double line and a corrugated “floor.” The purpose of the large illustration is to make the point that different engraving tools were used—a crucial step in the argument that this is not “decoration in a ‘geometric’ style.” The paragraph continues:
    
    This broad stroke begins the next series. Since the broad stroke occurs over the fine stroke, it must have been made later, and the full sequence on this face begins at the right and reads to the left, towards the hole. The sequence begins with an odd, delicately scratched sign, visible under the microscope (see [Plate 6]):
    
    The Plate reference here is to yet another detail, this time of the right–hand end of the bâton. This is the first moment that Marshack’s reconstructions have to be taken for granted, since in my copy of The Roots of Civilization the best that can be discerned from the photograph is something like this:
    
    but the marks are very indefinite in the photograph. The word “sign” does not appear in quotation marks this time, since the configuration is presented as the “odd” object that begins the first sequence; but by way of anticipating what he is about to do, he counts the nine component marks in the “odd, delicately scratched sign”:
    
    It is composed of 9 strokes. This is followed by a somewhat curved “Y” of 2 strokes, made by another point:
    
    Again the marginal picture in his text implicitly instructs readers about the limits of their participation in the decipherment, and about the limits of Marshack’s accuracy. The “Y” in Marshack’s photograph is clearly different, something more like this (Plate 6):
    
    At this point it is apparent, as it would already have been for a reader who had followed Roots of Civilization up to this point, that Marshack is relatively uninterested in the shape of the marks. I say relatively, since he is interested in shape insofar as it affects relative position— whether a mark is right or left of another, or above or below another in order of marking—and he is concerned with the shape of the floor of the mark itself. But the morphology of each mark is of less interest, because he is not going to be comparing configurations of overlapping or juxtaposed marks to other instances that might represent the same “sign” or “symbol.” Marshack continues:
    
    This is followed by a series of 14 strokes, made by still another point; the last ends with 2 small appended strokes:
    
    The regularity of the drawing of the 12 vertical strokes, as compared with the wobbly progress of the 12 marks as they appear in the accompanying photograph, makes it clear that Marshack’s drawing presents what he calls elsewhere a “schematic rendition” rather than an exact rendering. He uses the phrase for line drawings that employ different symbols (such as X or +) to discriminate between marks made with different burins, and he uses it in the captions that elucidate his proposed lunar sequences (48, 193). This kind of linear schematization is purposely normative, and it is intended to make an underlying pattern visible. What it does in this example is erase whatever subtleties of arrangement might also have been meaningful, in favor of a more–or–less regular parade of equal marks. Here, more clearly than before, the formal qualities of the bâton are being regimented in favor of the notational hypothesis.
    
    The next series, made by another point, has 1 appended stroke at the end. The full sequence on this face, with the counts by changes and differences of point, is shown in [Plate 8, top]. At the end of the series there are 10 strokes made by two points, in a 5 and 5 breakdown. Then, as though this fills out all the available notational space on this side of the bâton, over this, and between the last 5 long strokes, there are added 10 light strokes. These last 10 strokes place further restrictions on the reader’s participation in the analysis, since a look at Plate 5, top, shows something entirely different. I would give it schematically like this:
    
    The most important difference is that the two faint parallel lines in Marshack’s drawing (Plate 8) seem clearly to be continued upward through the X that he draws above them, and there are two very similar marks a little to the right that he nearly omits altogether. From the photograph it seems irrefutable that these two sets of parallel marks are made by single strokes, and their apparent doubleness is merely the double floor created by the burin that made the marks in the middle portion of the bâton. Several dozen other marks can be discerned in the photograph, all of them fugitive and dependent on the quality of reproduction in the book; but the sets of parallel lines seem to be as dependable as any form that he reproduces in his sketch. So again, in this case for the last time, the reader is warned away from too close a viewing, or too close a reading of Marshack’s text. It turns out this last discrepancy is the most important:
    
    All the marks on this face are intentional. It is obvious that this odd sequence of figures, counts, and groups is not ornamental or decorative, and that it is notational. Here as elsewhere in Roots of Civilization, the “ornamental” and “decorative” are disjoined from the “notational”: nothing can be both decorative and notational. Decoration is imagined as an activity with its own agenda and rules that are incompatible with notation. The word “intentional” has particular valence in Marshack’s writing, since he does not only mean it in a rigorous philosophic sense (as a mark interpreted to have been made with conscious purpose).
    
    He also intends something like “careful,” since presumably a decorative treatment of this bâton would have shown evidence of being more sloppily and quickly done, with one burin, at one time. Again the status of the “signs,” “figures” and “symbols” is in doubt, since there is an implication that the kind of care that went into making this notation is partly evidenced by the “signs”; but Marshack is about to compeltely ignore the signs in favor of merely counting their component marks.
    
    Is it lunar? Assuming one stroke per day, we lay the sequence against our lunar model [Plate 8, bottom]. The “lunar model” is the same throughout Roots of Civilization, since it functions as an astronomically correct schema against which “naked–eye observations of phases and periods” can be matched.26 Beneath the model Marshack places his schematic rendition of the marks on the bâton. The assumption, that one stroke equals one day, is consistent throughout the book, and never defended as anything more than a working hypothesis.27 In order to analyze the notation, both “signs” are dissected into component marks, one for each night’s appearance of the moon. He then concludes:
    
    This is an absolutely perfect tally for 2 months, beginning at last crescent and ending on the first day of invisibility 58 days later. We seem to have in this series a development of notational technique— recalling the accumulation of cueing marks of the Taï plaque [Plate 10]—in which the seeming “signs” or multiple marks, made by appended strokes of arms and groups of angles, are used to indicate points of change in lunar phase.
    Marshack’s reply to “On the Impossibility of Close Reading,” op. cit., p. [13]. In one place Marshack is particularly emphatic about the difference between his claim to have found notation and his hypothesis of the lunar calendar: “Of the hundreds of incised stones throughout Europe that I have studied in the last quarter of a century…, not one has provided evidence for a linear, sequetial ‘lunar’ notation.” Marshack, “On Wishful Thinking and Lunar ‘Calendars’” (reply to Francesco d’Errico), Current Anthropology 30 no. 4 (1989): 491.
    26 27
    
    How are they used? The lunar model reveals that on the day of an expected lunar observation around a phase point a small stroke is used, as though an observation of the moon was impossible on that day and the stroke represented a day of waiting. On the next clear day of observation the new series begins with a different stroking. Interesting also is the indication of a notational significance for the “quarter” moon, the visible half–moon. None of this is difficult to understand. It is once more a visual notational technique. It is non– arithmetical and it is cumulative. Remember also that the bâton was engraved some 5,000 years before agriculture formally “began” in the Fertile Crescent of the Middle East and some 10,000 years before the formal “beginning” of writing. The account of the bâton winds to a close with an invocation of two further engraved sequences, on the back and on one side: None of this is difficult to understand. It is once more a visual notational technique. It is non–arithmetical and it is cumulative. Remember also that the bâton was engraved some 5,000 years before agriculture formally “began” in the Fertile Crescent of the Middle East and some 10,000 years before the formal “beginning” of writing. On the reverse face of this bâton, scratched so faintly that even when held in hand most of the markings are invisible, yet clear under the microscope, is another unusual composition [Plate 9]. Once more there is certainty of notation, apparently for two lunar months. The lines, however, are so faintly scratched that a ballistic determination of point differences is difficult, and some exceedingly faint lines are difficult to prove.
    
    This Magdalenian IV bâton reveals a further stage in the development of the notational technique and a possible development in the case and precision of lunar observation and use. A third style of marking, a development of that style of parallel stroking overengraved to make a crosshatching, occurs along one side of this bâton. It, too, gives a sum for two months. The analysis ends in this way, with the puzzling “unusual composition” given a lunar phrasing of 27 + 30 days, and Marshack moves on to the next example.
    
    4 As Marshack himself has noted on several occasions, a first encounter with the “lunar theory” can leave readers incredulous. I want to try to capture some of that in order to exhibit the unexpected dynamic that leads from the reader’s certainty that Marshack is wrong, to a nagging uncertainty about the act of interpretation itself. It is this motion, I think, that accounts for the idea that Marshack is unanswerable or that he cannot be disproved. Each of the important difficulties with his theory can be written as a problem of distinguishing among the mobile lexicon of the rudiments of visual configurations: marks, icons, signs, schemata, notations, symbols, representations, and images. Over the course of his career Marshack has explored each of these, and they continue to be actively discussed in the field of Paleolithic archaeology; in this context I will mention just three.28
    Among the issues I will not be touching on here are symbols and “symbol systems.” See A. Marshack, “Upper Paleolithic Symbol Systems of the Russian Plain: Cognitive and Comparative Analysis,” Current Anthropology 20 no. 2 (1979): 271–94, with replies on 294–311. One of the letters responding to that essay, by Zbigniew Kobylínski and Urszula Kobylínska, tries to define “symbol” as a “representational sign that is semantically opaque, i.e., in the words of Roman Jakobson, has a poetic function…, or, as Soviet semioticians would have it, is self–reflective…; directly signifies sense perceptions…; is multivocal…; and has high capacity for modelling and evokes emotional–motivational states and therefore is perceived as identical with its referent….” (Ibid., 301; the ellipses are references in the text). The exchanges tend to run aground on theoretical imprecisions as much as on the reading of the artifacts themselves; but the topics always return eventually to what I am calling the rudiments of images. Another topic is the origin of representation; Marshack has claimed that the earliest representations are meanders, which he understands as a “‘point marker’ or symbol of change.” See Marshack, “The Meander as a System: The Analysis and Recognition of Iconographic Units in Upper Paleolithic Compositions,” in P. V. Ucko, editor, Form in Indigenous Art (Canberra: Australian Institute of Aboriginal Studies, 1977), 286–317. The
    28
    
    1. The problem of signs. An initial difficulty, well exemplified in the analysis of the Placard bâton, is that the “signs” and “symbols” have no role in the final lunar tally: instead they are picked apart into individual marks, and each mark is made equal to each other mark. At first that seems patently wrong: Why would a person arrange marks for appearances of the moon (or for any other phenomenon) into a carefully made, apparently intentional configuration, if all that were intended was a sequence of equal marks? It does not make sense, on the face of it, to propose that the meaning of the Placard bâton is a series of single marks, when it is comprised of collected and arranged “signs” of various types. The reasoning seems no more defensible than if Marshack had taken apart a print by Dürer into its individual lines. The apparent thoughtlessness of the analysis only seems worse when we consider what Marshack might reply. A reader of Roots of Civilization, for example, might expect him to say that he is merely following good scientific procedure and taking one hypothesis as far as it will go. Marshack himself points out the “signs” and their arrangements—the first is called an “odd, delicately scratched sign”—before he goes on with an analysis that does not take them into account. The problem here, if we are to judge by the criteria of a scientific hypothesis, is not that he chooses to ignore the “signs,” because no scientific theory need account for all phenomena presented by its sample—no close reading must be exhaustive. The “signs” have no meaning in the lunar schema, since both “signs” at the right of the bâton fall in the continuous period between the new moon and the first quarter. They may be meaningless, or they may have other meanings; Marshack could claim with perfect reason that they are not immediately under consideration. But that does not justify the way he excludes the “signs.” Marshack takes note of the relative positions of marks in two particular senses: he records how they cross over or under
    
    quotation is from a reply by Joel Gunn, Current Anthropology 20 no. 2 (1979): 299. Marshack has treated naturalistic images in another essay, “Some Implications of the Paleolithic Symbolic Evidence for the Origin of Language,” Current Anthropology 17 (1976): 274–82.
    
    one another, and he is interested in the sequence of marks, in this case from right to left. He does not consider relative positions of marks in any other sense (say, their relative positions when they comprise a “sign”). The “signs” are collections of individual marks made from right to left. But positionality is what defines the shapes of the “signs,” and so it does not seem logically sufficient to disassemble them into their component segments without justification. In addition Marshack does sometimes interpret position in a wider sense. He notes the “appended strokes of arms and groups of angles” of some marks, and shows how they fit the lunar sequence. A “small stroke,” he suggests, is used “on the day of an expected lunar observation around a phase point.” These observations have to do with the height, faintness, and spatial positions of marks. Why do they matter in these instances and not in the case of the “signs”? At this point it may seem that a critic would have a firm purchase on the analysis, both because marks are defined so oddly (in regard to some positions, and ignoring others) and also because the definition seems self–contradictory (since it variably excludes and includes spatial criteria such a size). If size and depth and overlap of mark can matter, then why not also marks that just touch, as in the “signs”? But there are deeper problems that stall this kind of objection. Marshack can appeal to the specificity of the artifact: who is to say the marks were not made in a way that entails exactly this definition of position? After all, we are considering a period before writing and before signs; perhaps the configurations of marks were meaningless to their makers in precisely these ways. Everything is recent, and nothing is native, where Paleolithic artifacts are concerned: there is no cultural link, no shared tradition of meaning or marking. An intractable aporia is produced by the act of identifying and then declining to read “signs,” since it must ultimately lead us to question the distinction between sign and mark, position–as–sequence and position–as–configuration. It is also the case that Marshack does interpret “signs” and other sets of non–notational marks in other notational contexts. His longest analysis—and one of the most intense and
    
    minute analyses of the semiotic elements of any visual object—concerns a 4 inch bone fragment from the Grotte du Taï scored with hundreds of tiny nicks and grooves (Plate 10). On that artifact Marshack finds notational marks—that is, the little hatches that appear here as vertical lines—and horizontal “containing lines” that serve to anchor the individual marks. But he also mentions “‘non–notational’ directional cueing marks,” defining the “direction and ‘flow’” of the notation (at the lower right), “an initiating, ‘non–notational’ symbolic motif,” perhaps representing “flow and motion” at the upper left, and “exceedingly fine engraved ‘hair’ lines” that “seem to mark the relevance of a day or period, but are distinct from the seqential accumulation of day units” (Marshack 1991, 28, 33, 47, 50, 60). Analytically, the most interesting moments in the account of the Taï plaque concern three further non–notational forms in the object’s upper–left corner, each one so small it is literally microscopic. The first is “a tiny horizontal containing line with 5 marks.” In Plate 11, it can be found to the right of the letter B and just inside the enclosure of heavy black lines. Another is “a tiny short sequence of 8 to 12 strokes between parallel lines (a ‘ladder’),” visible immediately to the right of the “tiny horizontal containing line.” Both of these, and especially the first, would be difficult to distinguish from the plaque’s other containing lines and notational marks, but Marshack says they were “clearly not part of the central notation” (Marshack 1991, 47). In the far upper–left corner, above the letter A, is a minuscule rectilinear meander “form” with six marks inside and three more on one corner—a configuration similar to the “signs” that Marshack declined to interpret, and determined to count, in the Placard bâton. Here the configuration is treated as a “form” that may have “positional relevance” but is again not integral to the notation and therefore need not be counted (Marshack 1991, 47). But why not? Because it is so much smaller than the other marks? But if so, then what determines the threshold of size beyond which a mark is no longer part of a notational sequence? The meander shape is also indistinguishably close to the “‘non–notational’ directional cueing marks” at the lower right of the artifact, and the notational marks often
    
    miss their containing lines as these marks do. By what criterion is this not part of the sequence? Before I draw a conclusion from these questions, it is worth noting that the answer cannot be historical. It is not sufficient to argue that “signs” of the type on the bâton are not likely to have meaning until the Neolithic, because at some point artifacts of the same age as the bâton will have had to have been analyzed and found deficient in “signs.” Conversely, when representational signs occur, for example the plant or “arrow” signs at Lascaux, their interpretation depends on assumptions about the internal rules of such images. Marshack offers only a quick look at the back of the Placard bâton, but if we linger we may begin to see the very “signs” and “pictures” that play such a large role later in Roots of Civilization, where they are interpreted as “seasonal,” “time–factored” images of plants. The little “sign” that looks like a bird’s foot on the back of the bâton is quite similar to a sign at Lascaux that Marshack reads as a “simplified branch,” in opposition to Leroi–Gourhan’s interpretation of it as a schematic “feminine” vulva–sign, or the Abbé Breuil’s reading of it as an arrow (223–24):
    
    Meaningless ÒsignÓ on the Placard b‰ton
    
    Branch in Lascaux
    
    Two configurations on the bâton that look like stalks (seventh and ninth from the left) are virtually identical to signs Marshack calls “plants” and “sprouts” in a later chapter (172, 216):
    
    ÒMeaninglessÓ configurations on the Placard b‰ton
    
    ÒPlantÓ on a b‰ton from Cueto de la Mina
    
    Marshack considers the problems of similar signs when it comes to Lascaux, since in that context he wants to show that plant signs on mobiliary artifacts might be close enough to some signs at Lascaux to imply they should be reinterpreted as plants (224). He does not reason that way about the Placard bâton because it is too early, but that does not explain why it is inappropriate to consider such possibilities in an analysis that presents itself as internal to the object.29 To return to the questions about when “signs” should be read: once again—and I am only taking isolated examples from a far larger field of possible issues—the objection about the status of “signs” on the Taï plaque seems cogent, but it can only succeed in raising the issue of assumptions. The “tiny horizontal containing line with 5 marks” at the upper left of the Taï artifact are not read as notation, and the fourteen lines are read together as a “form,” while the nine lines of the “odd, delicately scratched sign” at the right of the Placard bâton are read as notation. Each reading devolves from different assumptions about the possible relations of signs to individual marks. But what determines when marks are “executed differently” or may “represent a different category of marking”—as he says in another context, justifying the non–notational character of a set of marks—and therefore require a differing interpretation (Marshack 1991, 59)? Certainly the relation of one mark to others is essential in any asessment, but how is it possible to determine when the “position within the frame” is different enough to cross the border between mark and sign? Any and all objections
    For further development of this question of the influence of context on interpretation, see Marshack, “Reply,” Current Anthropology 20 no. 2 (1979): 303, where he argues that such images may elude modern distinctions between “representation, icon, sign, and symbol.”
    29
    
    of this kind, which are the only objections that are cogent within the terms of the text, must appear as rival hypotheses, and in that capacity they return to the questioner in the form of counter–questions about the reader’s axioms concerning the nature, relation, and disposition of marks. 2. The problem of notation. The lunar sequences have bothered readers of Roots of Civilization a little each time they are shown, because the fit and the principles of correspondence between marks and model differ with each succeeding example. Absolute accuracy is impossible and therefore meaningless, because as Marshack indicates the lunar cycle does not factor evenly into days, and measurement without instruments is always a matter of guesswork—the crescent moon cannot even be seen earlier than 15 hours after it is new, and there is a corresponding problem judging the hour of fullness. 30 But the absence of an absolute standard does not mean that relative accuracy cannot be measured, or that the principles of posited correspondence cannot be compared. Responding to Marshack’s analyses, Arnold R. Pilling objected that Marshack had failed to provide evidence that “any two objects have on them the same notational system.”31 Every calendar is the same, and all astronomers agree on the moon’s phases; hence it seems improbable that a “continent–wide system of notation” would not show more self– consistency.32 The assumption here is that notation of a single phenomenon involves detectibly uniform representation. But that is our own notion of notation, and it need not fit the Paleolithic one. In what sense are the lunar phases a single phenomenon, aside from our certainty that they are? To some readers, Marshack’s analyses seem to be mathematically indefensible since he accepts so many sequences as lunar cycles.33 By modern standards of
    M. Minnaert, The Nature of Light and Colour in the Open Air. Translated by H. M. Kremer-Priest, revision by K. E. Brian Jay (New York: Dover, 1954). 31 Arnold R. Pilling, Reply to “Cognitive Aspects of Upper Paleolithic Engraving,” Current Anthropology 13 no. 3–4 (1972), 469; Mary Aiken Littauer, “On Upper Paleolithic Engraving,” Anthropology Today 15 no. 3 (1974), 328. 32 G. F. Fry, Reply to A. Marshack, “Cognitive Aspects of Upper Paleolithic Engraving,” Current Anthropology 13 no. 3–4 (1972): 464. 33 Lewis–Williams and Dowson, “Theory and Data: A Brief Critique of A. Marshack’s Research Methods and Position on Upper Paleolithic Shamanism,” Rock Art Research 6 (1989):42–50, especially 49.
    30
    
    evidence, he is playing with numerological coincidences rather than requiring determinate ranges of fit—there is no mention of the usual scientific criteria for probabilistic matching, such as combined systematic and statistical error, quantified normative models or standard deviations. In terms of scientific method, there are often more marks in a given artifact that do not correspond to meaningful phase changes than marks that do, and more inconsistencies than parallels. Statistically speaking, Marshack’s hypotheses for objects such as the Placard bâton are a bad fit. Intermittently throughout Roots of Civilization Marshack says he tried many different correlations, and picked the best one, but the reader is left with the impression that others might be as good. In general, he claims to have done a great deal of empirical testing, which he has consistently not reported in his texts. There can be no doubt that he does perform such tests. In one passage, for example, a colleague mentions “[m]athematical, statistical, graphic, and sequential analysis of the results of… microscopic studies.”34 Still, in scientific terms these are serious, even crippling omissions. Marshack’s principal attempt to redress that again omits the discarded hypotheses and accounts of methodology that are standard in laboratory research.35 “Tests can be tried, within a range of possibilities, by anyone interested,” he remarks at one point.36 That is not entirely an interesting prospect, since if I were to try to rematch the lunar sequence, I would be reproving his hypothesis. Any rematch would be another instance of the same theory. There is no way to disprove the theory by adjusting the lunar sequence, and so the positioning comes to seem somewhat arbitrary. A reader of Roots of Civilization will tend, I think, to find each sequence about as plausible as every other—a scientifically untenable situation. But to all these questions, Marshack could ask in return what model, and what criteria, might be preferred in analyzing notations that are pre–arithmetical: the supposition is that they were made before there was such a thing as an arithmetical system that would allow the
    Hallam L. Movius, [Reply to Marshack], Current Anthropology 13 no. 3–4: (1972): 486. Marshack, “The Taï Plaque and Calendrical Notation in the Upper Paleolithic,” Cambridge Archaeological Journal 1 no. 1 (1991): 25–61. 36 Marshack, Roots of Civilization, op. cit. (1972 edition), 86.
    34 35
    
    summing of numbers or sets of numbers greater than, say, the range from one to ten.37 Wittgenstein thought of this problem, I think correctly, as an instance of the more general question of understanding when someone is following a rule. At one point in Remarks on the Foundation of Mathematics, he even imagines a “cave-man” who makes surprisingly ordered marks for no clear reason:
    
    There might be a cave–man who produced regular sequences of marks for himself. He amused himself, e.g., by drawing on the wall of the cave: —. ——. ——. ——. or —.—..—...—....— But he is not following the general expression of a rule. And when we say that he acts in a regular way that is not because we can form such and [such an] expression.38 For the present purposes, it is safe to ignore the phrase “he amused himself,” since Wittgenstein probably meant it as a reminder that the cave–man’s intentions and language are inaccessible. Wittgenstein also imagines a thrush that “always repeats the same phrase several times in a song.” Are we to conclude that “perhaps it gives itself a rule each time, and then follows the rule?”39 The stories go to show that regular actions need not follow regular rules, and that private “amusements” do not have rules in the ordinary sense. In the case of actual
    
    There is no necessary one–to–one correspondence between marks and days, so that in Marshack’s words, “Small sets of days could… be counted and marked at one moment and longer periods of observation would be marked each day without counting. This is the type of observation and counting common in many non– arithmeticalcalendars.” See Marshack’s reply to “On the Impossibility of Close Reading,” op. cit., p. [9].; and further Marshack, “The Chamula Calendar Board: An Internal and Comparative Analysis,” in Mesamerican Archaeology, New Approaches, edited by Norman Hammond (Austin, Texas: University of Texas Press: 1974): 255–70, especially 265; and also Marshack, “A Lunar–Solar Year Calendar Stick from North America,” American Antiquity 50 no. 1 (1985): 27–51, and Marshack, “North American Indian Calendar Sticks: The Evidence for a Widely Distributed Tradition,” in World Archaeoastronomy, edited by Anthony Aveni, pp. 308– 24. 38 Wittgenstein, Remarks on the Foundation of Mathematics, edited by G. H. von Wright, R. Rhees, and G. E. M. Anscombe (Cambridge, MA: MIT Press, 1983), 344. 39 Wittgenstein, Remarks on the Founbation of Mathematics, op. cit., 345.
    37
    
    Paleolithic notations, there is no mathematical standard available to use as a criterion of adequacy, since marks give no evidence of the marker’s conception of lunar phases—or to say it a little less exactly, the marks are the marker’s conception of notation. Should any notation be uniform, if notation itself was in its infancy? Should any fit be optimal, if the idea of representation itself was still inchoate? In this way cogent objections about Marshack’s lack of statistical rigor get turned into questions about the universality of quantification, numeration, and rigor itself. There have also been objections to Marshack’s reliance on tiny marks, at the threshold of invisibility.40 But why assume that the notations were referred to later, or that they could be read? Perhaps these notations were kinaesthetic: their meaning, as we would put it, was in their making. In that case any size that the burin could scratch would count—and as a corollary, the engraver might be indifferent to the size of the marks within the limits of the register lines or the available surface. Our thoughts about clarity and legibility—concepts that loom large in current texts on graphic communication—depend on the notion of notation as something that can be efficiently read: but why assume it should be read at all? Marshack is emphatic about the difference between his claim to have found notation and his hypothesis of the lunar calendar: “Of the hundreds of incised stones throughout Europe that I have studied in the last quarter of a century… not one has provided evidence for a linear, sequential ‘lunar’ notation.”41 If the claim and the hypothesis are kept separate, then the inadequacy of scientific objections to the lunar calendar becomes more obvious. The lunar calendar is susceptible to a number of critiques about methodology and falsification, but each of them has to appeal to a prior sense of notation. And what do we know about notation that might serve as a basis to argue against Marshack? Under what circumstances, with what evidence and what rules of interpretation, can we critique his assumptions?
    
    40 41
    
    Mary Aiken Littauer, “On Upper Paleolithic Engraving,” Anthropology Today 15 no. 3 (1974): 328. Marshack, “On Wishful Thinking and Lunar ‘Calendars’,” Current Anthropology 30 no. 4 (1989): 491.
    
    3. The problem of decoration. If “decoration” could be kept in quotation marks (as it often is in the recent literature) or defined negatively as non–utilitarian, non–symbolic, or non–notational marking, then there would be no special problem.42 But Marshack’s work seems to provoke the question of decoration because he sees so little of it in comparison to what other researchers would see. An artifact such as this pebble from Barma Grande (Plate 12) might seem like a formal experiment: perhaps the engraver was matching and contrasting lines with the shape of the pebble, and then quickly filling in blank areas with hatches and cross–hatches. It may look like what we expect of decoration; at any rate it does not correspond well to the structure of Marshack’s examples of Paleolithic notation, which are more one–dimensional even when the register lines snake back and forth. Yet in the first edition of Roots of Civilization he interprets it as notation, in which every mark has notational value. He surmises the longer parallel lines were intended “to give the engraver an enclosed or separate notational area”—that is, a one–dimensional track—“and therefore to serve as a visual and kinaesthetic aid.”43 He starts reading with the large curving set on the obverse (Plate 12, top), “since this central form or figure seems to take up the prime space on this face.” It is also possible, however, that the “angle–marks at top and bottom, serving almost as edge–marks,” might have been made first. The two choices are based on two ways of approaching the surface: either the person doing the marking began by taking charge of the blank surface, or by keeping to the edge. These are assumptions based partly on kinaesthetic hunches about ways of coming to terms with blank surfaces—and partly on assumptions about the ways images might be structured. The analysis moves forward with the discovery that every right–facing hatch mark within these register lines was made before every left– facing hatchmark; and so Marshack continues and eventually counts every line on the pebble. In the revised edition of Roots of Civilization the Barma Grande pebble is omitted, but the
    
    42 43
    
    Marshack, “On Wishful Thinking and Lunar ‘Calendars’,” op. cit., 491. Marshack, Roots of Civilization, first edition, op. cit., 84.
    
    question of reading developed two–dimensional patterns as notations remains.44 Under what circumstances, and with what criteria, could a more–or–less uniformly marked surface be interpreted as a sequence of individual notational marks? This more fully utilized surface example shows what kinds of problems are involved in determining sequences; Marshack generally says little about the trial–and–error process of finding the best order of reading for a surface thoroughly scored in two dimensions. Upper Magdalenian images show “the ability to synthesize and compose a “sum of images” into “storied” surfaces—in other words, to make meaningful narrative pictures—although the order of reading for such “non–linear series” is generally indeterminate (192, 211). What I am calling an “image” (what he calls, in this context, “decoration”) is something distinct from a notation: but the criteria of its difference are derived from very different kinds of artifacts. It seems Marshack is extending linear, one–dimensional configurations into two–dimensional multiples whose organizational laws cannot be derived from the one–dimensional cases. But what exactly are images, if not extensions of linear marks and linear accumulations of marks? And exactly how are images made, if not in this way? It has seemed to some archaeologists that Marshack does not recognize decoration, or allow for random, careless, and meaningless marking.45 In several papers, he has described about two dozen “symbol systems” including meanders, “fishlike images” (which resemble lattices or nets), zig–zags and double lines.46 But it can appear that he goes too far, and enlists the majority of decorations in the ranks of notations. Considering a bâton from Mezherich in the Ukraine, which is engraved with several sets of parallel hatchmarks, Marshack opts for notation over decoration even though the structure of the marking differs from the normative containing line and crossing hatchmark (Plate 13). It is relatively neutral to say the bâton has
    See for example Marshack, “The Taï Plaque and Calendrical Notation in the Upper Paleolithic,” op. cit., 148, 158, 216, 248, 254, 434, and especially Marshack, “Öküzini: Categorical Variation in the Symbolic Imagery,” in Fouiles à Öküzini: une site du Paléolithique final au sud de la Turquie, in press. 45 See for instance Francesco d’Errico, “Analyse technologique de l’art mobilier le cas de l’abri des Cabônes à Ranchot (Jura),” Galla Préhistoire 35 (1993): 139–76. 46 Marshack, “Upper Paleolithic Symbol Systems of the Russian Plain: Cognitive and Comparative Analysis,” Current Anthropology 20 no. 2 (1979): 271–95.
    44
    
    “une accumulation de séries de marques gravées selon des angles différents [an accumulation of sets of marks incised at different angles],” with “marques auxiliaires non structurées à l’extrême gauche [unstructured, subsidiary marks at the far left].” But the purpose of putting it that way is to emphasize the deliberation and internal structure of the marking: without actually counting the marks, he concludes that the bâton “représente un style personnel et idiosyncratique d’accumulation de séries de marques [represents a personal and idiosyncratic style of accumulating sets of marks].” It may represent “une accumulation mnémonique pour un rituel ou un mythe, ou la consignation de jours et d’événements [a mnemonic accumulation for a ritual or myth, or a record of days and events]” but either way, “la structure interne… suggère que la surface n’est pas «décorative» [the internal structure… suggests that this composition is not ‘decorative’].”47 To others, the markings on the bâton might appear sufficiently loose and unstructured so they could be classified as non–notational, if not intentionally “decorative.” “It seems that practically nothing is random,” one observer complains, “and very little is decorative.” 48 Others have objected that “more evolved ornamentation” can be “just as intentional, cumulative, and sequential” as some artifacts Marshack presents as notation.49 But where, Marshack might ask, do we get our ideas about the nature and frequency of decoration? Why can’t notation be the norm, and decoration be the uncommon counterexample? Marshack thinks that a “complex sequence of events, microscopically determined, eliminates the possibility of decorative intent or ‘art.’”50 But why would he not see the Mezherich bâton as a less carefully made artifact? Like the bâton, it looks like it might have been the work of a few minutes, and therefore, by our standards, likely to be “unthinking”
    Marshack, “l’Évolution et la transformation du décor du début de l’Aurignacien au Magdalénien final.” in L’Art des objets au Paléolithique, vol. 2, Les voies de la recherche, edited by Jean Clottes (Paris: Musées de France, 1987), 140–162, especially 149. See also Marshack, “Paleolithic Image,” in Encyclopedia of Human Evolution and Prehistory, edited by Ian Tattersall et al., 421–29, especially 427; and Marshack, [Reply to A. M. Byers], Current Anthropology 35 no. 4 (1994): 386–87. 48 Thomas Lynch, [Reply to Marshack], Current Anthropology 13 no. 3–4 (1972): 467. 49 Slavomil Vencl, [Reply to Marshack], Current Anthropology 13 no. 3–4 (1972): 470. 50 A. Marshack, “Reply” (to the critics of his “Cognitive Aspects of Upper Paleolithic Engraving,” op. cit.), Current Anthropology 13 no. 3–4 (1972): 474.
    47
    
    nonsemiotic marking or loose decoration.51 I can imagine decorating a strip between two register lines by making right–sloping marks after left–facing marks as in the Barma Grande pebble, or, as in the Mezherich bâton, by making a few sets of hatches, adding a register line, and moving on to the next set. But these are sliding scales: it is difficult not to see decoration as a kind of careless or immediate activity, as opposed to notation, and it is even more difficult to conceive notation as a potentially careless, quick, or relatively unstructured act. Decoration seems to entail a “regular design” made “rapidly and simply” and notation is somehow its opposite.52 In the critical literature, the question of the difference between notation and decoration has focused on Marshack’s claim to find frequent changes of marking tools within a single surface, making it more likely that the surfaces were marked in a deliberated, “time–factored” way. But that is a slippery point; the technical evidence for those claims is not clear, and the performative aspect of notations and decoration is too variable to make the division stick. A hand might change grip on a tool in the course of a few centimeters of marks; and a person recording a notation might put down a few marks quickly and carelessly, and then make the next few marks slowly and with a different tool.53 Francesco d’Errico and others have taken Marshack to task for his ways of distinguishing burins, and his arguments about the intervals between sequences of marks. Since d’Errico uses an electron microscope where Marshack uses an optical microscope, d’Errico’s readings are physically closer and he has been extracting different kinds of empirical data. For a while it seemed as if the problem of Paleolithic notation would come down to the atomic scale in its search for dependable criteria. d’Errico claims for example to have identified “les stries parasites,” the autographic marks of a single burin, so that he could tell definitively when a tool was re–used or when, as Marshack claims, one tool was allegedly exchanged for another in the course of a protracted series of
    See Whitney Davis, “The Origins of Image Making,” Current Anthropology 27 no. 3 (1989): 193–215. Francesco d’Errico, “Notation versus Decoration in the Upper Paleolithic: A Case–Study from Tossal de la Roca, Alicante, Spain,” Journal of Archaeological Science 21 no. 2 (1994): 185–200. 53 Marshack, [Reply], Current Anthropology 13 no. 3–4 (1972): 474.
    51 52
    
    temporal notations.54 But I would propose that the problem lies ultimately on an entirely different level: the concepts of decoration and notation are in some disarray, and they need to be addressed before any technical inquires can hope to make sense. d’Errico, for instance, uses a provisional tripartite classification of notations: those where “the elements constituting the system differ from one another,” those where “the elements are identical but can be differentiated by their distribution on the surface,” and those where “the marks cannot be differentiated” because they are identical and equidistant.55 But these three criteria apply to any number of “decorative” objects, and in particular they apply, respectively, to the bâton from Le Placard, the Mezherich bâton, and the Barma Grande pebble. By d’Errico’s working classification, each of them could be a notation. The question is not the analytical criteria for judging changes of tools, or the passage of time, or the slowness of the marking: it is the prior problem of the disheveled concepts that prompt those questions to begin with.
    
    5 It may seem that this has been an overly close examination of an obsessively close reader, but the ideas involved are widely applicable. Marshack’s method is not an eccentric or marginal regimen, but the vigilant application of a central methodology of any discipline that works with what it takes to be a structured, “systematic” artifact of any kind, from Paleolithic artifacts to Neoexpressionist canvases, and from graphs to literary texts.56 Marshack does what we all say we should do, but he does more of it, and in many respects he does it better: and for that reason the difficulty that people have experienced trying to argue with him needs to be taken seriously. Close readings are the best places to begin to ask what is meant by the most fundamental terms that are used to describe images. At least at the beginning, such an
    
    d’Errico, “Memoire et rythmes au Paléolithique: Le mythe des calendriers lunaires,” in Hominidæ, Proceedings of the Second International Congress of Human Paleontology, edited by Giacomo Giacobini (Milan, 1989), 507–10. 55 d’Errico, “Notation versus Decoration in the Upper Paleolithic,” op. cit.. 56 See Elkins, “Visual Schemata,” Encyclopedia of Aesthetics, edited by Michael Kelly, forthcoming.
    54
    
    inquiry could take a fairly rudimentary form, such as a simple list of assumptions about key terms. The list would have the virtue of turning conversations inward long enough to see what we expect in any given case of the central terms in our discourse about images. In regard to Marshack’s work, a provisional list of assumptions might include the following ideas about notation, some of them still unexamined in the literature: —Notation is comprised of marks disposed in a one–dimensional manner, either in a straight line, in rows, in curving lines, along an edge, or between register lines. If a surface is notational and is entirely covered with marks, it will break down into one–dimensional sequences. —Notation is “intentional,” meaning each mark is carefully and deliberately made. Sloppy or haphazard markings are more likely to be decorative, though they are often notational. (For example, there are ‘“sloppy’ ritual markings” such as ritual “killings” of images, and apparently “unstructured” accumulations of “‘structured’ motifs” can completely cover the surfaces of artifacts.57) —Decoration and notation are mutually exclusive, so that a marked artifact may be either decoration or notation, but not both. When a configuration of marks is “structural and sequential,” the “laws of probability argue against… an ‘artistic’ intention.” Decoration and “organized designs” do not usually show “unrhythmic crowding” of lines.58 —Notations are made over time, with measured pauses between each mark (they are “time–factored” in Marshack’s phrase). Marks may be made all at once, but if they are notational, they will represent an accumulation of time intervals that would stand for longer intervals. (This is not a rigid rule, since motifs, sets of marks, “‘signs’ such as hand prints, macaronis, ‘tectiforms’” and other “signs” might also be accumulated over time.59)
    
    For this point see Marshack’s reply to “On the Impossibility of Close Reading,” op. cit., p. [9]. Marshack, “The Taï Plaque,” op. cit., 36, 51 respectively. 59 Marshack, Roots of Civilization, op. cit., 1972 edition, 169 ff.; and Marshack’s reply to “On the Impossibility of Close Reading,” op. cit., p. [11].
    57 58
    
    —Notational sequences show “internal irregularities” or they divide into sets.60 Conversely, entirely uniform marks are likely to be decoration. —If a set of marks comprises a notation, every mark is significant and must be counted unless it is first classified as a cueing mark, register line, or other notational guide. —Notation does not make use of signs or symbols made of sets of marks. If a configuration is a notation, compound marks that appear to be signs or symbols need to be dissected into their individual marks.61 —The shape of a mark is generally not significant; it can be any size. Conversely, a shaped or representational mark will not normally be notational.62 (This an the next two points have important exceptions when it comes to cueing and other ancillary marks, and exceptions also occur where smaller marks correspond to phase points in the lunar sequence. I mean these final points in a statistical sense, since they describe what happens most frequently in Marshack’s interpretations as a whole, and what comprises the majority of his reading in any individual instance.) —The orientation of a mark is generally not significant; it can be horizontal or diagonal in a sequence that is primary vertical. Marks can be “feet” or complementary cross– hatching added to other marks that have already been laid down.63 —The position of a mark on a surface is generally not significant; a fully marked surface can sometimes be read by determining the order of marks, other times by trial and error.64 These are the kinds of questions that need to be addressed first, or along with, more practical or conventional interpretations, if the purpose is to increase our awareness of our own place in history and the expectations we have of artifacts (or to say it the other way
    
    Ibid., 86. Marshack, Roots of Civilization, op. cit., 62 Marshack, Roots of Civilization, op. cit., 63 Marshack, Roots of Civilization, op. cit., 64 Marshack, Roots of Civilization, op. cit.,
    60 61
    
    1991 edition, 39. 1972 edition, 230–31. 1991 edition, 149 ff. 1972 edition, 212.
    
    around, the meanings we “find” in them). That purpose is not always cogent, especially when interpretive communities are already posing and answering interesting questions; but when they become stalled with insoluble claims, as they have in Marshack’s case, these questions can dissolve some of the obstacles to further dialogue.
    
    6 One of the joys of Marshack’s work is that it shows just how loosely we read the more commonplace images that occupy most of anthropology and art history. Anthropologists interpret cultural artifacts in accord with assumptions about decoration that allow them to get on with the business of writing without stopping to count every welt in a scarified body, or every bead in a necklace. The best way to reveal those kinds of decisions is to contrast them with other interpretations based on different concepts of adequate or close reading. Scarifications and beadwork, after all, are “time–factored”: each bead takes a moment to string, and each piercing hurts. An interpretation that dissected a tattoo or a piece of jewelry into moments and gestures that fine might well appear as lunatic as Marshack’s lunar calendars have sometimes seemed: but that response disguises a serious problem—a hollowness in the concept of interpretation itself. The same is true in my own discipline. Art historians and critics pay little attention to groupings, types, sequences and series of marks, and virtually none to individual marks. Our smallest units tend to be images in their own right (depicted figures, portraits, objects, symbols) and when it comes to individual marks, as it does for example in Pollock’s work, we generally prefer theorizing about the nature of the marks to studying individual marks. All these are signs of what we overlook in order to preserve a certain working notion of close reading. By Marshack’s standards, our seeing is blurred and cursory, and our standards of looking are indefensibly lax. In literary criticism, where the concepts I have been exploring are most fully developed, the blindnesses of close readings affect the theory of close reading itself. Ginzburg is fully aware of his attraction to
    
    close reading in detective fiction, art history, and psychoanalysis, but only partly aware of the reasons why he is so attracted. His books replay that attraction by applying the method to various subjects, rather than delving into the critical question of the place of his writing within the tradition, and the limits of self–reflexivity in any given instance. Like art historians, anthropologists, and archaeologists, Ginzburg prefers close readings over looser ones, and he sometimes becomes interested in the history of close reading itself—but that interest does not intrude on his chosen regimens of close reading, which are unimpeded by any potentially disruptive critical reflection. Most of our readings are more or less—mostly more—lazy. Most of us want to escape the tyranny of prolonged close engagement with the artifacts or with our writing. Taking more ordinary encounters with objects as the standard, readings like Marshack’s (or Morelli’s) are apt to seem a little strange. Yet stiflingly close reading is an imperative of humanist scholarship and of literacy more generally: it has to be possible in order for there to be such a thing as a text, or an image, to understand; but it also has to be impossible, in order to let us get on with the vaguer meanings I think we prefer.
    
    * And of the five models I proposed at the beginning, which have been in play here? I think all of them have. I was drawn to Marshack’s work partly because I was interested in freeing myself from what I can now recognize as a version of iconography’s stultifying order (that is, my first model). The nearly microscopic scratches are clearly examples of things that are visually illegible, subsemiotic, “purely” visual, or nonsemantic (the second model). Marshack is very much the scientist, or scientific archaeologist (and so are several of his detractors), and in my somewhat dogged and often microscopic pursuit of the contradictions in his method, I am also behaving as a scientist (the third model). There is a poetry in these lost artifacts, which are seldom on display, have little immediate visual appeal, and are
    
    difficult to see in the original: they are, in fact, intensely Romantic (the fourth model). And finally, the entire enterprise is distinctly surrealist, and it owes its central insight to the Freudian model of repression — in fact, the entire essay would have been inconceivable without the advent and dissemination of surrealism and psychoanalysis (the fifth model). We are, as Derrida once said, still inside the age of psychoanalysis — and the period of modernism, and of science, and of Romanticism. Hence the obsessive interest in the detail.

Add Comment

To add a comment, you need to login or signup (you can signup without leaving this page)