This difference is relatively easy to demonstrate and the claim seems to have plausible potential. Yon Bard asked why we had to chose between weave pattern and closer descriptions and indeed we do not, excepting that there are few pictures of the backs of rugs in most rug books with good technical analyses. Marvin Amstey said he'd stay with the close descriptions, thank you, which led Wendel Swan to suggest that he might try Marvin's preferred system by not taking an photo ID to use to get past the guards at the Department of Labor but would take a verbal description of himself instead.

I thought that "weave pattern might be most useful in closer distinctions and offered four Turkmen pieces from Neff and Maggs, woven by different tribes and also four Hammadans from a 1993 catalog published by the Swedish dealer, J. P. Willborg, who is one of the few people to have followed Neff and Maggs, lead, by including life-size photos of the backs of the 43 rugs in this exhibition and sale. (Pat Seiler subsequently pointed out that Brian MacDonald has also included photos of rug backs in his recent "Tribal Rugs.")

In a side conversation Marla Mallett seemed to agree that weave pattern was likely to be more useful in closer distinctions but also added the caveat that she expected that it would be most useful for attributing rugs woven after 1880 in factory-like settings, since most tribal weaving had more than one weaver and was not subjected to the pressures for uniformity at the level at which these likely existed in commercial endeavors.

But when we began to examine rugs things seemed to turn out differently. Vincent Keers put up a Hammadan and asked whether we could make a closer attribution using weave pattern. I found at least four similar backs in Willborg (Wendel said that there are several more possibles) but when we examined the closer technical descriptions that both Vincent and Willborg provided, things broke down and we were unable ultimately to offer a closer attribution for Vincent's piece.

Weave pattern seemed to work better on rugs at the extremes. I offered a yastik with what I'm pretty sure are goat hair warps and dark brown wefts, as one rug I have with a back I think I could recognize anywhere. Later in the salon I offered a second Caucasian-seeming piece which has a remarkable amount of white on the back. Vincent offered two quite similar backs of pieces of different ages that at a great distance from each other by weavers of different ethic groups. We couldn't tell much without additional information. But Wendel put up three verbal technical descriptions without any images or additional information to demonstrate that, by themselves, these don't work very well either.

One seemingly telling instance was our examination at their invitation of an "old Sarouk" and a "new Sarouk," provided by Neff and Maggs explicitly to show the utility of weave pattern in making a distinction that might slip by close verbal description. But Vincent wondered whether their new Sarouk might not be a Jozan and Wendel said that their "old Sarouk" might be as well. These are disconcerting suggestions since they mark what seems potentially to be an attribution mistake Neff and Maggs have made themselves while demonstrating the utility of the weave pattern perspective. This mistake would be unlikely at the level of close description, since Sarouks have and asymmetric knot and Jozans have symmetric ones.

The fact that we obviously use a perspective in our everyday lives, as we recognize family and friends, that is seemingly close to what Neff and Maggs are recommending, led to a discussion of some facets of perception. Steve Price and others noted that it is remarkable how accurately we can recognize our children and our associates and even someone we've seen only rarely and often we can do this on the basis of only partial "looks" or even bad images. This in turn led to comment about how loosely weave pattern seems to be used, how little effort is made to test the assertions of experts using it, and how difficult it might in fact be to detect the extent to which it is being used, since it is almost always used in conjunction with other information about the rug in question. We tend to look at the front and THEN at the back. So we almost always have quite a bit of additional information when we do so.

Towards the end of this part of the discussion I mused about "tacit" knowledge, whether it is primarily "unstated" or perhaps at some levels "unstate-able." Richard Farber suggested at this point that our discussion had implications for how children should be introduced to art and to music.

Finally, our wondering about why more rug books with good technical information have not provided life-size images of the rugs backs, led to a discussion of economical ways in which this could be done. Ken Thompson talked about his personal experience using CD-ROM to distribute large amounts of information and testified that the costs are very low. Steve Price and I talked about the need for good search capabilities with CD-ROM distributions. And a subsequent face-to-face discussion with one of the Turkmen collectors who are providing us with material for the in-hotel Turkmen exhibition at ICOC X revealed that, because they have 700-800 images of high quality Turkmen and other Central Asian pieces, and because the book they plan can accommodate only about 200 of these, they are actively exploring whether the balance of this material might not be published via CD-ROM.

I think we ended less impressed with Neff and Maggs' notion of weave pattern than the initial plausibility of it and our experience with the accuracy of our everyday perceptions suggested it might be. But it still seems to me that Neff and Maggs have noticed something that is potentially important and that perhaps rug books that want to offer comprehensive technical information are mistake to leave it out. We do not appear to be able to use it at the level of precision with which we operate as we distinguish our children but there is something going on at the level of weave pattern that may be worthy of more rigorous efforts than those we have made to date.

Thanks to those who contributed to, what I think may well have been, an interesting discussion.

Regards,

R. John Howe