4 The Application of the Quantitative Paradigm to Historical Sociolinguistics: Problems with the Generalizability Principle JUAN M. HERNÁNDEZ-CAMPOY AND NATALIE SCHILLING Historical sociolinguistics has often been considered to suffer, perhaps inevitably, from lack of representativeness and validity of its findings. This is because the sociolinguistic study of historical language forms must rely on linguistic records from previous periods – most of which will be incomplete or non-representative in some way – as well as on knowledge and understanding of past sociocultural situations that can only be reconstructed rather than directly observed or experi- enced by the researcher. In this paper, we mention the seven main problems we have to contend with when trying to practice historical sociolinguistic research: i) representativeness, ii) empirical validity, iii) invariation, iv) authenticity, v) author- ship, vi) social and historical validity, and vii) standard ideology. But theoretical and procedural problems are also present even in the apparently rigorous method- ology of variationist sociolinguistics, as evidenced, for example, in the unevenness with which studies conform or fail to conform to the generalizability principle, as well as the extent to which ‘generalizability’ in sociolinguistic studies is even pos- sible. Indeed, as Bailey and Tillery (2004: 13) note, “without a body of research that examines the effects of methods on results,” we cannot even know how generalizable (i.e. representative and reliable) our studies have been or may be. Hence, we cannot hold historical sociolinguistics to standards with which The Handbook of Historical Sociolinguistics, First Edition. Edited by Juan Manuel Hernández-Campoy, Juan Camilo Conde-Silvestre. © 2012 Blackwell Publishing Ltd. Published 2012 by Blackwell Publishing Ltd.