Meaning and Mining: the Impact of Implicit Assumptions in Data Mining for the Humanities

@article{citeulike:1363834, abstract = {In working across and between disciplines, it is the tacit assumptions that may be most destructive to meaningful collaboration. Ours is a state of mutual ignorance, and the goals and practice of the professional literary historian and the machine-learning researcher are equally obscure. But in collaboration mutual ignorance becomes an opportunity for self-reflection, clarification, and the speaking of what is usually unspoken. Willard McCarty writes, “Computational form, which accepts only that which can be told explicitly and precisely” proves “useful for isolating ... tacit and inchoate” knowledge (256). Collaborators are forced to set out a program in detail, one that is mutually comprehensible but also one that delivers results that are simultaneously meaningful in two disciplines. In this paper, we discuss the tacit assumptions that accompany data set preparation, hypothesis testing, and data exploration in order to deliver prescriptive claims. We propose a communication protocol designed to bring hidden and tacit assumptions into plain view where they may be discussed and analyzed. This paper is the third in a series of collaborative efforts undertaken by the two authors. It is informed by real experience working together: working often at cross purposes, garbling a common language, but ultimately producinresults that are of interest to both computer scientists and literary scholars.}, author = {Sculley, D. and Pasanek, Brad }, citeulike-article-id = {1363834}, journal = {Digital Humanities 2007}, keywords = {collaboration, computer-science, humanities, literary-computing, machine\_learning}, month = {June}, posted-at = {2007-06-04 22:33:18}, priority = {0}, title = {Meaning and Mining: the Impact of Implicit Assumptions in Data Mining for the Humanities}, url = {http://www.digitalhumanities.org/dh2007/abstracts/xhtml.xq?id=224}, year = {2007} }

TY - JOUR ID - citeulike:1363834 L3 - citeulike-article-id:1363834 TI - Meaning and Mining: the Impact of Implicit Assumptions in Data Mining for the Humanities JF - Digital Humanities 2007 N2 - In working across and between disciplines, it is the tacit assumptions that may be most destructive to meaningful collaboration. Ours is a state of mutual ignorance, and the goals and practice of the professional literary historian and the machine-learning researcher are equally obscure. But in collaboration mutual ignorance becomes an opportunity for self-reflection, clarification, and the speaking of what is usually unspoken. Willard McCarty writes, “Computational form, which accepts only that which can be told explicitly and precisely” proves “useful for isolating ... tacit and inchoate” knowledge (256). Collaborators are forced to set out a program in detail, one that is mutually comprehensible but also one that delivers results that are simultaneously meaningful in two disciplines. In this paper, we discuss the tacit assumptions that accompany data set preparation, hypothesis testing, and data exploration in order to deliver prescriptive claims. We propose a communication protocol designed to bring hidden and tacit assumptions into plain view where they may be discussed and analyzed. This paper is the third in a series of collaborative efforts undertaken by the two authors. It is informed by real experience working together: working often at cross purposes, garbling a common language, but ultimately producinresults that are of interest to both computer scientists and literary scholars. KW - collaboration KW - computer-science KW - humanities KW - literary-computing KW - machine_learning AU - Sculley, D AU - Pasanek, Brad PY - 2007/06// UR - http://www.digitalhumanities.org/dh2007/abstracts/xhtml.xq?id=224 ER -