THE NON-DIAGNOSTIC SHERD PROBLEM: LESSONS FROM JEDDITO YELLOW WARE TYPOLOGY

Abstract

Although ceramic types are applied most commonly to sherds (simply because of their abundance), most typologies have been derived at least in part from patterns of attribute covariation on whole vessels. As a result, one typological concern is the “non-diagnostic sherd problem,” where only a fraction of the sherds deriving from a whole vessel share all of the attributes of that vessel. This paper assesses the severity of the non-diagnostic sherd problem for Jeddito Yellow Ware polychrome sherds, demonstrating that it may cause up to half of the sherds from some polychrome vessels to be incorrectly typed. Implications for chronology and social dynamics are discussed, and a preliminary correction for the non-diagnostic sherd problem is offered.

A pesar de que las clases cerámicas frecuentemente se aplican a tiestos (sobre todo por razón de su abundancia), la mayoría de las tipologías se derivan en parte de atributos concurrentes en vasijas enteras. Así queda la cuestión de “el problema de tiestos no-diagnósticos”, en lo cual sólo una minoría de los tiestos de una vasija manifiestan todos los atributos de la vasija. El presente artículo valora la gravedad de “el problema de tiestos no-diagnósticos” acerca de los tiestos polícromos de Jeddito Yellow Ware, y demuestra que el problema ya mencionado resulta en una clasificación equivocada de hasta la mitad de las piezas polícromas. Las consecuencias en fechar la cerámica son tratadas y propone una corrección preliminar.

ACKNOWLEDGEMENTS

I am grateful to Steven LeBlanc, Bob Edberg, and Kelley Hays-Gilpin for their invaluable assistance with the Yellow Ware Online Database. Helpful comments on drafts of this paper were provided by Linda Cordell, Charles Adams, Kelley Hays-Gilpin, and an anonymous reviewer. I also thank the participants of the 2009 Yellow Ware Ceramic Conference at the Museum of Northern Arizona and the Amerind Seminar on Jeddito Yellow Ware in 2008 for sharing their knowledge of JYW typology.

Notes

1 Temper density in Jeddito Yellow Ware was first studied by Smith (1971:478–480), and while he did find that heavily-tempered pieces are most common in the early fourteenth-century assemblages (Awatovi Black-on-yellow) and temperless pieces are typical of assemblages from the late fourteenth century and after (Jeddito Black-on-yellow and Sikyatki Polychrome), there is significant overlap in temper density in the transitional period. Up to 1/3 of the ostensibly temperless Jeddito Black-on-yellow and Sikyatki Polychrome vessels exhibited heavy temper densities typical of Awatovi Black-on-yellow, while 1/4 of the typically heavily tempered Awatovi Black-on-yellow vessels exhibited light temper densities expected of Jeddito Black-on-yellow and Sikyatki Polychrome. LaMotta (2006) confirmed the general temper trend in a study of Jeddito Yellow Ware from Homol'ovi I and Homol'ovi II, but made use of the pattern only at the level of the assemblage, not for the purposes of typing individual sherds. Thus, while there is clearly a directional continuum of temper density over the fourteenth century, the fuzziness of the pattern makes temper a relatively unreliable marker for typing individual sherds.

Likewise, Smith's (1971:476–477) study of Jeddito Yellow Ware paste color documented a general trend of darker earlier colors (in the Munsell range 7.5YR to 10YR) and lighter later colors (1oYR to 2.5YR), but significant variability: 30 percent of Awatovi Black-on-yellow vessels were “white”, while 25 percent of Jeddito Black-on-yellow vessels were reddish brown.

There remains debate among Jeddito Yellow Ware ceramic specialists about whether the variability around temper and paste color trends is sufficient to exclude them from a sherd typology or whether, in combination with other time-sensitive attributes, they can be used profitably.

2 Paayu Polychrome, which contains thinned shades of black or brown in massed areas to create the effect of a polychrome, dates from 1365 to 1385 in deposits at Homol'ovi I and II and is a clear predecessor to Sikyatki Polychrome. This type has only recently been recognized (Hays 1991) and was not distinguished in any of the analyses used here.

3 An Early Sikyatki Polychrome variant (ca. A.D. 1385–1400) is also recognized, featuring geometric designs in which red paint is used to outline black solid areas, but this variant is very rare and is not considered in this study.

4 The American Museum of Natural History, the Arizona State Museum, the Field Museum, the Brooklyn Museum of Art, the Denver Museum of Natural History, the Logan Museum, the Heard Museum, the Maxwell Museum, the Southwest Museum, the Museum of Northern Arizona, the Peabody Museum, the Smithsonian Museum, the Phoenix Museum of Natural History, the University of Colorado Museum, the Museum of New Mexico, Montezuma Castle National Monument, the University of Pennsylvania Museum of Archaeology and Anthropology.

5 Type production spans were assumed to follow a Gaussian distribution with an alpha parameter of 2. Bichrome production dates were set to: introduced A.D. 1300, middle of production A.D. 1400, terminated A.D. 1500. Sikyatki Polychrome production dates were set to: introduced A.D. 1385, middle of production A.D. 1500, terminated A.D. 1630.