Language distributions from EEBO to EarlyPrint Douglas Knox Language distributions from EEBO to EarlyPrint EarlyPrint is a representation of early modern English-language texts. EarlyPrint is essentially a subset of the texts transcribed by the Text Creation Partnership, and it leaves out almost all TCP texts in languages other than English. The TCP itself was a curated selection from the works with images in EEBO, which has its origins in a microfilming project that drew from long years of bibliographic work that began with the Pollard and Redgrave and Wing bibliographies, among others, culminating in what is now the English Short Title Catalog. The original scope of the bibliographies sought to capture works printed in England in any language as well as works printed in English anywhere in the world during a particular time span. At each stage of curation from ESTC to EEBO to TCP to EarlyPrint, there has been a steady exclusion of works that are primarily not in English. This is convenient for building a corpus to model the English language and for applying computational tools to provide linguistic annotation. Many English-language EarlyPrint works still of course contain occasional passages both long and short that are in other languages. MorphAdorner part-of-speech tags do not attempt to characterize parts of speech in other languages, but do identify words as “foreign” and attempt to identify the language, such as “fla” for Latin, “ffr” for French, etc. One might want to be cautious about understanding EarlyPrint as representative of early modern English print culture generally, especially insofar as linguistic diversity was a part of that culture. The table below shows the distribution of language attributions from EEBO metadata across EEBO, TCP, and EarlyPrint. (“Newari” is undoubtedly a data error, perhaps from a stage of data entry where an attribute value ‘new’ had some other meaning.) This data does not illuminate other relevant questions one might consider, including whether works available to be catalogued had differential survival rates by language, genre, and audience, as well as the nature of the international book trade and the circulation within England of non-English-language works imported from elsewhere. Publication Language in EEBO in TCP in EP % in TCP % in EP English 135,132 59,340 58,809 43.9 43.5 Latin 9,069 514 10 5.7 0.1 Romance (Other) 795 11 0 1.4 0 French 679 102 5 15 0.7 Welsh 205 150 0 73.2 0 Ancient Greek 147 2 0 1.4 0 Dutch 140 13 0 9.3 0 Greek 112 0 0 0 0 Italian 95 17 0 17.9 0 Scots 88 39 1 44.3 1.1 French, Middle (ca. 1400-1600) 80 15 0 18.8 0 Undetermined 78 2 0 2.6 0 Multiple languages 73 6 0 8.2 0 German 63 3 0 4.8 0 Spanish 32 6 0 18.8 0 Not applicable 31 5 2 16.1 6.5 Irish 19 1 0 5.3 0 Algonquin 12 0 0 0 0 Hebrew 12 1 0 8.3 0 North American Indian (other) 9 1 0 11.1 0 [blank] 8 3 3 37.5 37.5 Gaelic (Scots) 8 5 0 62.5 0 Portuguese 7 1 0 14.3 0 Arabic 6 0 0 0 0 Newari 5 1 1 20 20 English, Old (ca. 450-1100) 5 0 0 0 0 French, Old (ca. 842-1400) 4 1 0 25 0 Malay 1 0 0 0 0 Hungarian 1 0 0 0 0 Syriac 1 0 0 0 0 Russian 1 0 0 0 0 Turkish 1 0 0 0 0 Coptic 1 0 0 0 0 Persian 1 0 0 0 0 Lithuanian 1 0 0 0 0