The Literary Surface: A Plea for the Use of Measurement in Narratological Analysis Kim Jautze Huygens ING kim.jautze@huygens.knaw.nl Roel Smeets Utrecht University R.J.H.Smeets@uu.nl Generally, analyses of literary representation are pursued on the basis of close readings of a confined amount of texts. In this mode of analysis, central narratological concepts as focalization are applied to literary cases in order to gain insight into (ideological) structures of representation (e.g. “oŶg ϮϬϭ5; MiŶŶaaƌd ϮϬϭϬͿ. “tudies oŶ, foƌ iŶstaŶĐe, geŶdeƌ hieƌaƌĐhies iŶ liteƌatuƌe utilize Bals (2009) distinction between a focalizing subject and a focalized object (e.g, Buikema 2009, Meijer 1996). However, the inevitable intuitive and subjective nature of close readings obstructs the generalizability of such interpretations. In order to achieve a more intersubjective and generalizable interpretation, the computer can provide repeatable and verifiable methods to examine a large corpus of texts, thereby generating a more empirically solid basis. Digital tools, Moretti (2013) argues, may provide us with the opportunity to gain new insights into ever existing classifications in literary studies. Narratological concepts and hypotheses are being examined in the field of computational narratology. Whereas studies on superficial textual features (e.g. stylistic elements as the use of words or bigrams) do not necessarily require in-depth computational models, analysis of narrative features do require such models. In order to automatically distract character types without any prior knowledge requires building intelligent models (Bamman et al. 2014). For such studies it is necessary to have had a solid education in computer science, which most of the traditionally educated literary scholars do not haǀe. It is theƌefoƌe iŵagiŶaďle that theLJ feaƌ the ƌise of ĐoŵputatioŶal liteƌaƌLJ aŶalLJsis. Underlying such resistance towards Digital Humanities could be 1) the difficulty of applying distant- reading approaches without being educated in programming and statistics, as well as 2) a general criticism on the reliability of computational measurements. 1 ElaďoƌatiŶg oŶ that, ǁe addƌess tǁo ƋuestioŶs: ϭͿ aƌe theƌe loǁ-leǀel textual features, i.e. features ǁhiĐh lie oŶ the tedžts suƌfaĐe, 2 which can be examined with a basal knowledge of computers? And if so, 2) how reliable are the outcomes of such analyses? On the basis of the answers to these 1 In this paper we do not address the arguments relating to the concern how literary scholars in Digital Humanities aim to come to objective measurements, which, in the opinion of most traditionalist, is in contradiction with the subjective nature of literary interpretations. 2 Best aŶd MaƌĐus ;ϮϬϬϵͿ defiŶe the suƌfaĐe as ǁhat is eǀideŶt, peƌĐeptiďle, appƌeheŶsiďle iŶ tedžts […]. A surface is what insists on being looked at ƌatheƌ thaŶ ǁhat ǁe ŵust tƌaiŶ ouƌselǀes to see thƌough ;p.ϵͿ.