Language Extraction from Zinc Sulphide A Dissertation Presented for the Doctor of Philosophy Degree University of Tennessee, Knoxville Dowman Parks Varn December 2001 Abstract Recent advances in the analysis of one-dimensional temporal and spacial series allow for detailed characterization of disorder and computation in physical systems. One such system that has defied theoretical understanding since its discovery in 1912 is polytypism. Polytypes are layered compounds, exhibiting crystallinity in two dimensions, yet having complicated stacking sequences in the third direction. They can show both ordered and disordered sequences, sometimes each in the same specimen. We demonstrate a method for extracting two-layer correlation information from ZnS diffraction patterns and employ a novel technique for epsilon-machine reconstruction. We solve a long-standing problem -- that of determining structural information for disordered materials from their diffraction patterns -- for this special class of disorder. Our solution offers the most complete possible statistical description of the disorder. Furthermore, from our reconstructed epsilon-machines we find the effective range of the interlayer interaction in these materials, as well as the configurational energy of both ordered and disordered specimens. Finally, we can determine the `language' (in terms of the Chomsky Hierarchy) these small rocks speak, and we find that regular languages are sufficient to describe them. A copy of this dissertation in pdf format: Language Extraction from ZnS Citation: D.P. Varn, Language Extraction from ZnS, Ph.D. Thesis, University of Tennessee, Knoxville, (2001).