After a BA in Linguistics at the University of Essex, Tamsin Blaxter did her MPhil in General Linguistics and Comparative Philology at the University of Oxford, with a dissertation that explored gendered language in the medieval Íslendingasögur. For her PhD at the University of Cambridge, she worked on language change and space in medieval Norway. Her work focuses on new methods of gathering and exploring large datasets in variationist linguistics, with a particular interest in space.

• List of figuresList of tablesAbbreviations and symbolsAcknowledgements1. Introduction 1.1. Foreword1.2. Structure2. The statistics of space in historical linguistics 2.1. Introduction2.1.1. Visualising time and space: problems and solutions2.2. Kernel smoothing2.2.1. Kernel smoothing for spatial variation2.2.2. Time2.3. Parameter setting2.3.1. Setting bandwidths based on real-world priors2.3.2. Setting bandwidths based on predictive accuracy2.3.3. Setting bandwidths based on power analysis2.3.4. Conclusions3. Language in medieval Norway and sources of data 3.1. Introduction3.2. Norway and Norwegian in the late medieval period3.2.1. Sociolinguistic statuses of the languages3.2.2. Contact3.2.3. A note on Jamtland3.3. The Diplomatarium Norvegicum and the DN online3.3.1. The Diplomatarium Norvegicum3.3.2. Annotation4. Variable definitions and datasets 4.1. Introduction4.2. Phonological variables4.2.1. Rise of svarabhakti vowels4.2.2. Loss of voiceless dental fricatives4.2.3. Loss of voiced dental fricatives4.3. Morphological variables4.3.1. Changes in agreement of verbs with 1sg. subjects4.3.2. Changes in the form of 1pl and 1du. subject pronouns4.3.3. Changes in agreement of verbs with 1pl and 1du. subjects4.4. Syntactic variables4.4.1. Loss of lexical genitives5. Individual changes 5.1. Introduction5.2. The loss of the dental fricatives5.2.1. Timing5.2.2. Spatial distribution5.3. The loss of 1pl active -um5.3.1. Timing5.3.2. Spatial distribution5.4. The Constant Rate Effect5.5. Conclusions6. The Trudgill conjecture 6.1. Introduction6.2. Theoretical background6.2.1. Trudgill, Dahl and McWhorter6.2.2. Other proposals regarding relationships between complexity and social factors6.2.3. The case of Nordic6.3. Predictions6.4. Variables6.5. Spatial distributions6.5.1. Loss of 1sg agreement in the present active6.5.2. Loss of the voiced dental fricative6.5.3. The rise of svarabhakti vowels6.5.4. Loss of the voiceless dental fricative6.5.5. Merger of 1du. pronoun into 1pl pronoun6.5.6. Merger of 1pl pronoun into 1du. pronoun6.5.7. Replacement of the 1pl pronoun by borrowed vi6.5.8. Replacement of the 1du. pronoun by borrowed vi6.5.9. Loss of 1pl middle agreement6.5.10. Loss of genitive objects of millum6.5.11. Loss of 1pl agreement in verb-subject order6.5.12. Loss of 1pl agreement in subject-verb order6.5.13. Loss of number agreement on present tense verbs6.6. Conclusions7. Spatial models of diffusion 7.1. Background7.2. Methods7.2.1. Approach7.2.2. Hierarchicality measure7.3. Variables7.3.1. Replacement of 1pl v´er by reanalysed m´er7.3.2. Replacement of 1du. vit by reanalysed mit7.4. Results7.5. Discussion8. Conclusions 9. Annex: A step-by-step guide to kernel smoothing for historical dialectology9.1. Introduction9.2. Software and dependencies9.3. Preparing your data9.4. Preliminary exploration9.5. Determining bandwidths9.6. Temporal kernel smoothing9.7. Spatial-temporal kernel smoothing9.8. Spatial visualisation10. Appendix

After a BA in Linguistics at the University of Essex, Tamsin Blaxter did her MPhil in General Linguistics and Comparative Philology at the University of Oxford, with a dissertation that explored gendered language in the medieval Íslendingasögur. For her PhD at the University of Cambridge, she worked on language change and space in medieval Norway. Her work focuses on new methods of gathering and exploring large datasets in variationist linguistics, with a particular interest in space.

• List of figuresList of tablesAbbreviations and symbolsAcknowledgements1. Introduction 1.1. Foreword1.2. Structure2. The statistics of space in historical linguistics 2.1. Introduction2.1.1. Visualising time and space: problems and solutions2.2. Kernel smoothing2.2.1. Kernel smoothing for spatial variation2.2.2. Time2.3. Parameter setting2.3.1. Setting bandwidths based on real-world priors2.3.2. Setting bandwidths based on predictive accuracy2.3.3. Setting bandwidths based on power analysis2.3.4. Conclusions3. Language in medieval Norway and sources of data 3.1. Introduction3.2. Norway and Norwegian in the late medieval period3.2.1. Sociolinguistic statuses of the languages3.2.2. Contact3.2.3. A note on Jamtland3.3. The Diplomatarium Norvegicum and the DN online3.3.1. The Diplomatarium Norvegicum3.3.2. Annotation4. Variable definitions and datasets 4.1. Introduction4.2. Phonological variables4.2.1. Rise of svarabhakti vowels4.2.2. Loss of voiceless dental fricatives4.2.3. Loss of voiced dental fricatives4.3. Morphological variables4.3.1. Changes in agreement of verbs with 1sg. subjects4.3.2. Changes in the form of 1pl and 1du. subject pronouns4.3.3. Changes in agreement of verbs with 1pl and 1du. subjects4.4. Syntactic variables4.4.1. Loss of lexical genitives5. Individual changes 5.1. Introduction5.2. The loss of the dental fricatives5.2.1. Timing5.2.2. Spatial distribution5.3. The loss of 1pl active -um5.3.1. Timing5.3.2. Spatial distribution5.4. The Constant Rate Effect5.5. Conclusions6. The Trudgill conjecture 6.1. Introduction6.2. Theoretical background6.2.1. Trudgill, Dahl and McWhorter6.2.2. Other proposals regarding relationships between complexity and social factors6.2.3. The case of Nordic6.3. Predictions6.4. Variables6.5. Spatial distributions6.5.1. Loss of 1sg agreement in the present active6.5.2. Loss of the voiced dental fricative6.5.3. The rise of svarabhakti vowels6.5.4. Loss of the voiceless dental fricative6.5.5. Merger of 1du. pronoun into 1pl pronoun6.5.6. Merger of 1pl pronoun into 1du. pronoun6.5.7. Replacement of the 1pl pronoun by borrowed vi6.5.8. Replacement of the 1du. pronoun by borrowed vi6.5.9. Loss of 1pl middle agreement6.5.10. Loss of genitive objects of millum6.5.11. Loss of 1pl agreement in verb-subject order6.5.12. Loss of 1pl agreement in subject-verb order6.5.13. Loss of number agreement on present tense verbs6.6. Conclusions7. Spatial models of diffusion 7.1. Background7.2. Methods7.2.1. Approach7.2.2. Hierarchicality measure7.3. Variables7.3.1. Replacement of 1pl v´er by reanalysed m´er7.3.2. Replacement of 1du. vit by reanalysed mit7.4. Results7.5. Discussion8. Conclusions 9. Annex: A step-by-step guide to kernel smoothing for historical dialectology9.1. Introduction9.2. Software and dependencies9.3. Preparing your data9.4. Preliminary exploration9.5. Determining bandwidths9.6. Temporal kernel smoothing9.7. Spatial-temporal kernel smoothing9.8. Spatial visualisation10. Appendix

After a BA in Linguistics at the University of Essex, Tamsin Blaxter did her MPhil in General Linguistics and Comparative Philology at the University of Oxford, with a dissertation that explored gendered language in the medieval Íslendingasögur. For her PhD at the University of Cambridge, she worked on language change and space in medieval Norway. Her work focuses on new methods of gathering and exploring large datasets in variationist linguistics, with a particular interest in space.

• List of figuresList of tablesAbbreviations and symbolsAcknowledgements1. Introduction 1.1. Foreword1.2. Structure2. The statistics of space in historical linguistics 2.1. Introduction2.1.1. Visualising time and space: problems and solutions2.2. Kernel smoothing2.2.1. Kernel smoothing for spatial variation2.2.2. Time2.3. Parameter setting2.3.1. Setting bandwidths based on real-world priors2.3.2. Setting bandwidths based on predictive accuracy2.3.3. Setting bandwidths based on power analysis2.3.4. Conclusions3. Language in medieval Norway and sources of data 3.1. Introduction3.2. Norway and Norwegian in the late medieval period3.2.1. Sociolinguistic statuses of the languages3.2.2. Contact3.2.3. A note on Jamtland3.3. The Diplomatarium Norvegicum and the DN online3.3.1. The Diplomatarium Norvegicum3.3.2. Annotation4. Variable definitions and datasets 4.1. Introduction4.2. Phonological variables4.2.1. Rise of svarabhakti vowels4.2.2. Loss of voiceless dental fricatives4.2.3. Loss of voiced dental fricatives4.3. Morphological variables4.3.1. Changes in agreement of verbs with 1sg. subjects4.3.2. Changes in the form of 1pl and 1du. subject pronouns4.3.3. Changes in agreement of verbs with 1pl and 1du. subjects4.4. Syntactic variables4.4.1. Loss of lexical genitives5. Individual changes 5.1. Introduction5.2. The loss of the dental fricatives5.2.1. Timing5.2.2. Spatial distribution5.3. The loss of 1pl active -um5.3.1. Timing5.3.2. Spatial distribution5.4. The Constant Rate Effect5.5. Conclusions6. The Trudgill conjecture 6.1. Introduction6.2. Theoretical background6.2.1. Trudgill, Dahl and McWhorter6.2.2. Other proposals regarding relationships between complexity and social factors6.2.3. The case of Nordic6.3. Predictions6.4. Variables6.5. Spatial distributions6.5.1. Loss of 1sg agreement in the present active6.5.2. Loss of the voiced dental fricative6.5.3. The rise of svarabhakti vowels6.5.4. Loss of the voiceless dental fricative6.5.5. Merger of 1du. pronoun into 1pl pronoun6.5.6. Merger of 1pl pronoun into 1du. pronoun6.5.7. Replacement of the 1pl pronoun by borrowed vi6.5.8. Replacement of the 1du. pronoun by borrowed vi6.5.9. Loss of 1pl middle agreement6.5.10. Loss of genitive objects of millum6.5.11. Loss of 1pl agreement in verb-subject order6.5.12. Loss of 1pl agreement in subject-verb order6.5.13. Loss of number agreement on present tense verbs6.6. Conclusions7. Spatial models of diffusion 7.1. Background7.2. Methods7.2.1. Approach7.2.2. Hierarchicality measure7.3. Variables7.3.1. Replacement of 1pl v´er by reanalysed m´er7.3.2. Replacement of 1du. vit by reanalysed mit7.4. Results7.5. Discussion8. Conclusions 9. Annex: A step-by-step guide to kernel smoothing for historical dialectology9.1. Introduction9.2. Software and dependencies9.3. Preparing your data9.4. Preliminary exploration9.5. Determining bandwidths9.6. Temporal kernel smoothing9.7. Spatial-temporal kernel smoothing9.8. Spatial visualisation10. Appendix