Computer-Assisted Language Comparison ... and its Application to Chinese Dialectology

Slide 1

Slide 1 text

Computer-Assisted Language Comparison... ... and its Application to Chinese Dialectology Johann-Mattis List DFG Research Fellow Centre des recherches linguistiques sur l’Asie Orientale Team AIRE (Adaptation, Integration, Reticulation, Evolution) at UPMC 2015-02-04 1 / 50

Slide 2

Slide 2 text

Murder on the Orient Express… 2 / 50

Slide 3

Slide 3 text

2 / 50

Slide 4

Slide 4 text

- 1 victim (dead) - 12 stab wounds - all stab wounds diﬀer from each other - 12 suspects - all have an alibi 2 / 50

Slide 5

Slide 5 text

- 1 victim (dead) - 12 stab wounds - all stab wounds diﬀer from each other - 12 suspects - all have an alibi ? 2 / 50

Slide 6

Slide 6 text

Eh oui, Hastings, it seems to me that there is only one possible soluti- on, no matter how impossible it may seem: There was not one murderer, but twelve… 2 / 50

Slide 7

Slide 7 text

Well, why not? My reasoning has always been based on the assumption that when you have ex- cluded the impossible, whatever re- mains, however im- probable, must be the truth. 2 / 50

Slide 8

Slide 8 text

Criminology Historical Linguistics Goal ﬁnd the murderer ﬁnd the proto-language Proceeds by reconstructing the reconstructing the circumstances of a crime history of a language Method inference based on inference based on circumstantial evidence circumstantial evidence Mode of Reasoning abduction abduction 2 / 50

Slide 9

Slide 9 text

Criminology Historical Linguistics *ʔˁak 2 / 50

Slide 10

Slide 10 text

Mode of Reasoning in Historical Linguistics The starting point of historical linguistics is syn- chronic language data. Based on models of language and sound change, an inference of relations among diﬀerent data points is carried out. An analysis of these relations yields historical scena- rios that explain the current structure of the data. 2 / 50

Slide 11

Slide 11 text

Slide 12

Slide 12 text

Traditional Historical Linguistics 3 / 50

Slide 13

Slide 13 text

Traditional Historical Linguistics Characteristics Characteristics FRANZ BOPP VERY, VERY LONG TITLE 4 / 50

Slide 14

Slide 14 text

Traditional Historical Linguistics Characteristics Research Object FRANZ BOPP VERY, VERY LONG TITLE 5 / 50

Slide 15

Slide 15 text

Traditional Historical Linguistics Characteristics Research Object FRANZ BOPP VERY, VERY LONG TITLE German ʦ aː n - * Proto-Germanic t a n d English t ʊː θ - ** Proto-Indo-European d o n t Italian d ɛ n t e * Proto-Romance d e n t French d ɑ̃ - - 5 / 50

Slide 16

Slide 16 text

Slide 17

Slide 17 text

Traditional Historical Linguistics Characteristics Research Object FRANZ BOPP VERY, VERY LONG TITLE German ʦ aː n - - * Proto-Germanic t a n d English t ʊː - θ - ** Proto-Indo-European d o n t Italian d ɛ n t e * Proto-Romance d e n t French d ɑ̃ - - - 5 / 50

Slide 18

Slide 18 text

Traditional Historical Linguistics Characteristics Research Object FRANZ BOPP VERY, VERY LONG TITLE German ʦ aː n - - Proto-Germanic t a n θ - English t ʊː - θ - ** Proto-Indo-European d o n t Italian d ɛ n t e Proto-Romance d e n t e French d ɑ̃ - - - 5 / 50

Slide 19

Slide 19 text

Traditional Historical Linguistics Characteristics Research Object FRANZ BOPP VERY, VERY LONG TITLE German ʦ aː n - Proto-Germanic t a n θ - English t ʊː - θ ** Proto-Indo-European d o n t Italian d ɛ n t e Proto-Romance d e n t e French d ɑ̃ - - 5 / 50

Slide 20

Slide 20 text

Traditional Historical Linguistics Characteristics Research Object FRANZ BOPP VERY, VERY LONG TITLE German ʦ aː n - Proto-Germanic t a n θ - English t ʊː - θ Proto-Indo-European d e n t - Italian d ɛ n t ə Proto-Romance d e n t e French d ɑ̃ - - 5 / 50

Slide 21

Slide 21 text

Traditional Historical Linguistics Characteristics Research Object FRANZ BOPP VERY, VERY LONG TITLE German ʦ aː n - * Proto-Germanic t a n d English t ʊː - θ Proto-Indo-European d e n t Italian d ɛ n t ə * Proto-Romance d e n t French d ɑ̃ - - 5 / 50

Slide 22

Slide 22 text

Traditional Historical Linguistics Characteristics Research Object FRANZ BOPP VERY, VERY LONG TITLE German ʦ aː n Proto-Germanic t a n θ English t ʊː θ Proto-Indo-European d e n t Italian d ɛ n t e Proto-Romance d e n t e French d ɑ̃ German ʦ aː n Proto-Germanic t a n θ English t ʊː θ Proto-Indo-European d e n t Italian d ɛ n t e Proto-Romance d e n t e French d ɑ̃ 1 5 / 50

Slide 23

Slide 23 text

Traditional Historical Linguistics Characteristics Research Object FRANZ BOPP VERY, VERY LONG TITLE History individual events (description) individual processes (description, inference) general processes (modeling, analysis) Language History individual language states (description of sound system, grammar, lexicon) individual instances of language development (description and inference of sound change patterns, grammaticalization, lexical change) general language development (modeling and analysis of sound change processes, grammaticalization, lexical change) 6 / 50

Slide 24

Slide 24 text

Traditional Historical Linguistics Characteristics Research Object FRANZ BOPP VERY, VERY LONG TITLE Internal Language History (ontogenesis) etymology historical grammar historical phonology External Language History (phylogenesis) linguistic reconstruction (modeling and inference) proof of language relationship (inference) genetic classiﬁcation (analysis) General Tendencies in Language History processes and mechanisms of sound change grammaticalization lexical change 7 / 50

Slide 25

Slide 25 text

Traditional Historical Linguistics Characteristics Origins FRANZ BOPP VERY, VERY LONG TITLE Uniformitarianism “universality of change” – change is independent of time and space “graduality of change” – change is neither abrupt nor chaotic “uniformity of change” – change is not heterogeneous, but uniform Founding Fathers Franz Bopp (1791–1867): language comparison (Bopp 1816) Rasmus Rask (1787-1832) and Jacob Grimm (1785-1863): sound law (Rask 1818, Grimm 1822) August Schleicher (1821–1868): family tree and linguistic reconstruction (Schleicher 1853 & 1861) 8 / 50

Slide 26

Slide 26 text

Traditional Historical Linguistics Achievements Achievements 9 / 50

Slide 27

Slide 27 text

Traditional Historical Linguistics Achievements Methods and Theories Comparative Method (Meillet 1925) Basic procedure for proving language relationship and reconstructing unattested ancestral language states, etymologies, and genetic classiﬁcations. Family Tree Model and Wave Theory (Schleicher 1853, Schmidt 1872) Two partially incompatible models to describe historical language relations. Regularity Hypothesis (Osthoﬀ & Brugmann 1878) Fundamental working hypothesis that states that certain sound change processes proceed regularly (universally, gradually, and in a uniform manner). 10 / 50

Slide 28

Slide 28 text

Traditional Historical Linguistics Achievements Comparative Method proof of relationship identification of cognates identification of sound correspondences reconstruction of proto-forms internal classification 11 / 50

Slide 29

Slide 29 text

Slide 30

Slide 30 text

Traditional Historical Linguistics Achievements Comparative Method Cognate List Alignment Correspondence List German dünn d ʏ n GER ENG Frequ. d θ 3 x d d 1 x n n 1 x m m 1 x ŋ ŋ 1 x English thin θ ɪ n German Ding d ɪ ŋ English thing θ ɪ ŋ German dumm d ʊ m English dumb d ʌ m German Dorn d ɔɐ n English thorn d ɔː n 11 / 50

Slide 31

Slide 31 text

Slide 32

Slide 32 text

Traditional Historical Linguistics Achievements Comparative Method Cognate List Alignment Correspondence List German dünn d ʏ n GER ENG Frequ. d θ 2 x d d 1 x n n 1 x m m 1 x ŋ ŋ 1 x English thin θ ɪ n German Ding d ɪ ŋ English thing θ ɪ ŋ German dumm d ʊ m English dumb d ʌ m German Dorn d ɔɐ n English thorn d ɔː n 11 / 50

Slide 33

Slide 33 text

Slide 34

Slide 34 text

Traditional Historical Linguistics Achievements Comparative Method Cognate List Alignment Correspondence List German dünn d ʏ n GER ENG Frequ. d θ 3 x d d 1 x ? n n 2 x m m 1 x ŋ ŋ 1 x English thin θ ɪ n German Ding d ɪ ŋ English thing θ ɪ ŋ German dumm d ʊ m English dumb d ʌ m German Dorn d ɔɐ n English thorn θ ɔː n 11 / 50

Slide 35

Slide 35 text

Traditional Historical Linguistics Achievements Comparative Method Cognate List Alignment Correspondence List German dünn d ʏ n GER ENG Frequ. d θ 3 x d d 1 x n n 2 x m m 1 x ŋ ŋ 1 x English thin θ ɪ n German Ding d ɪ ŋ English thing θ ɪ ŋ German dumm d ʊ m English dumb d ʌ m German Dorn d ɔɐ n English thorn θ ɔː n 11 / 50

Slide 36

Slide 36 text

Traditional Historical Linguistics Achievements Insights Internal Language History Thanks to historical linguistics, the history of a considerable (but still small) amount of languages has been thoroughly investigated. External Language History Thanks to historical linguistics, a considerable amount of the languages in the world has been genetically classiﬁed (although there remain many unsolved and controversially discussed questions). General Language History Some work on general processes of language history has been done, yet many questions still remain unsolved or are controversially debated. 12 / 50

Slide 37

Slide 37 text

Traditional Historical Linguistics Problems Problems 13 / 50

Slide 38

Slide 38 text

Traditional Historical Linguistics Problems Transparency 14 / 50

Slide 39

Slide 39 text

Traditional Historical Linguistics Problems Transparency Part of the process of “becoming” a competent Indo-Europeanist has always been recognized as coming to grasp “intuitively” concepts and types of changes in language so as to be able to pick and choose between alternative explanations for the history and development of speciﬁc features of the reconstructed language and its oﬀspring. Schwink (1994) 14 / 50

Slide 40

Slide 40 text

Traditional Historical Linguistics Problems Applicability 15 / 50

Slide 41

Slide 41 text

Traditional Historical Linguistics Problems Applicability – 7,106 languages (Lewis & Fennig 2013) – 147 language families (ibid.) – 25244065 languages which could be compared 15 / 50

Slide 42

Slide 42 text

Traditional Historical Linguistics Problems Applicability The amount of digitally available data for the languages of the world is growing from day to day, while there are only a few historical linguists who are trained to carry out the comparison of these languages. It seems impossible to handle this task when relying only on the traditional, time- consuming manual procedures developed in traditional historical linguistics. 15 / 50

Slide 43

Slide 43 text

Traditional Historical Linguistics Problems Adequacy 16 / 50

Slide 44

Slide 44 text

Traditional Historical Linguistics Problems Adequacy One time is never, two times is ever! (a mathematician friend on the treatment of probability in Indo-European linguistics) 16 / 50

Slide 45

Slide 45 text

Traditional Historical Linguistics Problems Comparability 17 / 50

Slide 46

Slide 46 text

Traditional Historical Linguistics Problems Comparability Frucht. Sf std. (9. Jh.), mhd. vruht, ahd. fruht, as. fruht. Ent- lehnt aus l. frūctus m. gleicher Bedeutung (zu l. fruī “genie- ße”). Das deutsche Wort ist Femininum geworden im An- schluß an die ti- Abstrakta wie Flucht² usw. Adjektive: fruch- tig, fruchtbar; Verb: (be-)fruchten. Ebenso nndl. vrucht, ne. fruit, nfrz. fruit, nschw. frukt, nnorw. frukt; frugal. (Kluge und Seebold 2002) 17 / 50

Slide 47

Slide 47 text

Traditional Historical Linguistics Problems Comparability inherited from borrowed from derived from PIE *bhreu ̯ Hg ̑ - “to use” PIE *bhruHg ̑ -ié- “to use” (present tense) PGM *ƀrūkan- “to use” OHG brūhhan “to use” G brauchen “to use” G Brauch “custom” OHG fruht “profit, fruit” G frugal “nourishing” Fr fruit “profit,fruit” Fr frugal “modest (food)” Lt fruor, fruī “I enjoy” Lt frūctus “profit” Lt frux “fruit, grain” Lt frugalis “bring profit” 17 / 50

Slide 48

Slide 48 text

Traditional Historical Linguistics Problems Summary Aspect Current State Desired State transparency intuition formalization applicability diﬃcult and slow fast enough to analyze our heritage before we loose it adequacy lack of probabilistic reasoning inclusion of probabilistic arguments comparability philological, irreproducible knowledge representation formal, digital, and reproducible knowledge representation 18 / 50

Slide 49

Slide 49 text

Quantitative Historical Linguistics 19 / 50

Slide 50

Slide 50 text

Quantitative Historical Linguistics Characteristics Characteristics P(A|B)=(P(B|A)P(A))/(P(B) 20 / 50

Slide 51

Slide 51 text

Quantitative Historical Linguistics Characteristics Characteristics P(A|B)=(P(B|A)P(A))/(P(B) “Indo-European and computational cladistics” (Ringe, Warnow and Taylor 2002) “Language-tree divergence times support the Anatolian theory of Indo-European origin” (Gray und Atkinson 2003) “Language classiﬁcation by numbers” (McMahon und McMahon 2005) “Curious Parallels and Curious Connections: Phylogenetic Thinking in Biology and Historical Linguistics” (Atkinson und Gray 2005) “Automated classiﬁcation of the world’s languages” (Brown et al. 2008) “Computational Feature-Sensitive Reconstruction of Language Relationships: Developing the ALINE Distance for Comparative Historical Linguistic Reconstruction” (Downey et al. 2008) “Networks uncover hidden lexical borrowing in Indo-European language evolution” (Nelson-Sathi et al. 2011) “A pipeline for computational historical linguistics” (Steiner, Stadler, und Cysouw 2011) 21 / 50

Slide 52

Slide 52 text

Slide 53

Slide 53 text

Quantitative Historical Linguistics Characteristics Points of Interest and Goals P(A|B)=(P(B|A)P(A))/(P(B) phylogenetic reconstruction sequence comparison general questions of language development Goals If we cannot guarantee getting the same results from the same data considered by diﬀerent linguists, we jeopardize the essential scientiﬁc criterion of repeatability. (McMahon & McMahon 2005) 22 / 50

Slide 54

Slide 54 text

Quantitative Historical Linguistics Characteristics Methods and Theories P(A|B)=(P(B|A)P(A))/(P(B) phylogenetic reconstruction (cf., among others, Gray & Atkinson 2003 Ringe et al. 2002, Brown et al. 2008) phonetic alignment (cf., among others, Kondrak 2000, Prokić et al. 2009, List 2012a) cognate detection (cf. Steiner et al. 2011, List 2012b) borrowing detection (cf. Nelson-Sathi et al. 2011, List et al. 2014a) 23 / 50

Slide 55

Slide 55 text

Quantitative Historical Linguistics Achievements Achievements 24 / 50

Slide 56

Slide 56 text

Quantitative Historical Linguistics Achievements New Perspectives external language history receives more attention than before “Indo-Euro-Centrism” is replaced by a more cross-linguistic paradigm new questions regarding general language history new proposals to model language history 25 / 50

Slide 57

Slide 57 text

Quantitative Historical Linguistics Achievements New Approaches empirical data becomes the center of interest probabilistic approaches replace “historical” approaches databases replace philological knowledge representation “informal” methods are formalized and automatized 26 / 50

Slide 58

Slide 58 text

Quantitative Historical Linguistics Achievements Examples 27 / 50

Slide 59

Slide 59 text

Quantitative Historical Linguistics Achievements Examples: Phonetic Alignment Alignment Analyses Alignment analyses display sequence similarities by representing multiple sequences as rows of a matrix in which common segments are placed in the same column. Alignments are a formal way to deal with general tasks of sequence comparison. Although never explicitly labeled or displayed, alignments are virtually present in all analyses in historical linguistics dealing with the comparison of sound sequences (words, morphemes). 27 / 50

Slide 60

Slide 60 text

Slide 61

Slide 61 text

Slide 62

Slide 62 text

Slide 63

Slide 63 text

Quantitative Historical Linguistics Achievements Examples: Phonetic Alignment Sound Classes Sounds which frequently occur in correspondence relations in genetically related languages can be clustered into classes. It is thereby assumed that “phonetic correspondences inside a ‘type’ are more regular than those between diﬀerent ‘types’” (Dolgopolsky 1986[1964]: 35). 28 / 50

Slide 64

Slide 64 text

Slide 65

Slide 65 text

Slide 66

Slide 66 text

Slide 67

Slide 67 text

Slide 68

Slide 68 text

Slide 69

Slide 69 text

Quantitative Historical Linguistics Achievements Examples: Phonetic Alignment Sound-Class-Based Phonetic Alignment (SCA, List 2012 a) Sound classes and alignment analyses can be combined. Sound sequences are internally represented as sound classes. Alignments are carried out using standard algorithms developed in evolutionary biology. INPUT tɔxtər dɔːtər TOKENIZATION t, ɔ, x, t, ə, r d, ɔː, t, ə, r CONVERSION t ɔ x … → T O G … d ɔː t … → T O T … ALIGNMENT T O G T E R T O - T E R CONVERSION T O G … → t ɔ x … T O - … → d oː - … OUTPUT t ɔ x t ə r d ɔː - t ə r 1 29 / 50

Slide 70

Slide 70 text

Quantitative Historical Linguistics Achievements Examples: Automatic Cognate Detection LexStat (List 2012, List 2014) LexStat is a method for automatic cognate detection in multilingual wordlists. It uses on sound-class-based sequence alignment (SCA) analyses as a proxy to infer language-speciﬁc sound similarities (similar to the notion of sound correspondences in historical linguistics). Using the automatically inferred sound similarities, LexStat partitions words into cognate sets. 30 / 50

Slide 71

Slide 71 text

Quantitative Historical Linguistics Achievements Examples: Automatic Cognate Detection Cognate Clustering Analysis ID Taxa Word Gloss GlossID IPA ... ... ... ... ... ... 21 German Frau woman 20 frau 22 Dutch vrouw woman 20 vrɑu 23 English woman woman 20 wʊmən 24 Danish kvinde woman 20 kvenə 25 Swedish kvinna woman 20 kviːna 26 Norwegian kvine woman 20 kʋinə ... ... ... ... ... ... 30 / 50

Slide 72

Slide 72 text

Quantitative Historical Linguistics Achievements Examples: Automatic Cognate Detection Cognate Clustering Swedish English Danish Norwegian Dutch German kvinna woman kvinde kvine vrouw Frau Swedish kvina 0.00 0.69 0.07 0.12 0.71 0.78 English wumin 0.69 0.00 0.66 0.57 0.68 0.87 Danish kveni 0.07 0.66 0.00 0.08 0.67 0.71 Norwegian kwini 0.12 0.57 0.08 0.00 0.75 0.74 Dutch frou 0.71 0.68 0.67 0.75 0.00 0.17 German frau 0.78 0.87 0.71 0.74 0.17 0.00 Analysis ID Taxa Word Gloss GlossID IPA ... ... ... ... ... ... 21 German Frau woman 20 frau 22 Dutch vrouw woman 20 vrɑu 23 English woman woman 20 wʊmən 24 Danish kvinde woman 20 kvenə 25 Swedish kvinna woman 20 kviːna 26 Norwegian kvine woman 20 kʋinə ... ... ... ... ... ... 30 / 50

Slide 73

Slide 73 text

Quantitative Historical Linguistics Achievements Examples: Automatic Cognate Detection Cognate Clustering Swedish English Danish Norwegian Dutch German kvinna woman kvinde kvine vrouw Frau Swedish kvina 0.00 0.69 0.07 0.12 0.71 0.78 English wumin 0.69 0.00 0.66 0.57 0.68 0.87 Danish kveni 0.07 0.66 0.00 0.08 0.67 0.71 Norwegian kwini 0.12 0.57 0.08 0.00 0.75 0.74 Dutch frou 0.71 0.68 0.67 0.75 0.00 0.17 German frau 0.78 0.87 0.71 0.74 0.17 0.00 German Frau frau Dutch vrouw vrou English woman wumin Danish kvinde kveni Swedish kvinna kvina Norwegian kvine kwini 30 / 50

Slide 74

Slide 74 text

Quantitative Historical Linguistics Achievements Examples: Automatic Cognate Detection Cognate Clustering Swedish English Danish Norwegian Dutch German kvinna woman kvinde kvine vrouw Frau Swedish kvina 0.00 0.69 0.07 0.12 0.71 0.78 English wumin 0.69 0.00 0.66 0.57 0.68 0.87 Danish kveni 0.07 0.66 0.00 0.08 0.67 0.71 Norwegian kwini 0.12 0.57 0.08 0.00 0.75 0.74 Dutch frou 0.71 0.68 0.67 0.75 0.00 0.17 German frau 0.78 0.87 0.71 0.74 0.17 0.00 German Frau frau Dutch vrouw vrou English woman wumin Danish kvinde kveni Swedish kvinna kvina Norwegian kvine kwini 30 / 50

Slide 75

Slide 75 text

Quantitative Historical Linguistics Achievements Examples: Automatic Cognate Detection Cognate Clustering German Frau frau Dutch vrouw vrou English woman wumin Danish kvinde kveni Swedish kvinna kvina Norwegian kvine kwini Analysis ID Taxa Word Gloss GlossID IPA CogID ... ... ... ... ... ... ... 21 German Frau woman 20 frau 1 22 Dutch vrouw woman 20 vrɑu 1 23 English woman woman 20 wʊmən 2 24 Danish kvinde woman 20 kvenə 3 25 Swedish kvinna woman 20 kviːna 3 26 Norwegian kvine woman 20 kʋinə 3 ... ... ... ... ... ... ... 30 / 50

Slide 76

Slide 76 text

Quantitative Historical Linguistics Achievements Examples: Automatic Cognate Detection INPUT TOKENIZATION PREPROCESSING LOG-ODDS D ISTANCE COGNATE OUTPUT CORRESPONDENCE DETECTION USING PHONETIC ALIGNMENT LOOP DISTRIBUTION LexStat Algorithm (List 2014) EXPECTED ATTESTED DISTRIBUTION CALCULATION CLUSTERING 30 / 50

Slide 77

Slide 77 text

Quantitative Historical Linguistics Achievements Examples: Automatic Cognate Detection B-Cubed F-Scores on BDCD Benchmark (List 2014) Bai (Tibeto-Burman) Indo-European Japanese and Ryukyu Ob-Ugrian Austronesian Sinitic (Chinese Dialects) 60 65 70 75 80 85 90 95 Turchin NED SCA LexStat 31 / 50

Slide 78

Slide 78 text

Slide 79

Slide 79 text

Slide 80

Slide 80 text

Quantitative Historical Linguistics Achievements Examples: Phylogenetic Networks Hugo Schuchardt (1842-1927) 32 / 50

Slide 81

Slide 81 text

Quantitative Historical Linguistics Achievements Examples: Phylogenetic Networks Hugo Schuchardt (1842-1927) “We connect the branches and twigs of the tree with countless horizon- tal lines and it ceases to be a tree.” (Schuchardt 1870 [1900]: 11) 32 / 50

Slide 82

Slide 82 text

Quantitative Historical Linguistics Achievements Examples: Phylogenetic Networks Hugo Schuchardt (1842-1927) 32 / 50

Slide 83

Slide 83 text

Quantitative Historical Linguistics Achievements Examples: Phylogenetic Networks Hugo Schuchardt (1842-1927) 32 / 50

Slide 84

Slide 84 text

Quantitative Historical Linguistics Achievements Examples: Phylogenetic Networks 33 / 50

Slide 85

Slide 85 text

Quantitative Historical Linguistics Achievements Examples: Phylogenetic Networks -- Spanish -- French -- Italian Danish -- English -- German -- 33 / 50

Slide 86

Slide 86 text

Quantitative Historical Linguistics Achievements Examples: Phylogenetic Networks -- Spanish -- French -- Italian Danish -- English -- German -- 33 / 50

Slide 87

Slide 87 text

Quantitative Historical Linguistics Achievements Examples: Phylogenetic Networks -- Spanish -- French -- Italian Danish -- English -- German -- 33 / 50

Slide 88

Slide 88 text

Quantitative Historical Linguistics Achievements Examples: Phylogenetic Networks -- Spanish -- French -- Italian Danish -- English -- German -- 33 / 50

Slide 89

Slide 89 text

Quantitative Historical Linguistics Achievements Examples: Phylogenetic Networks List et al. (2014b) . . ---Lánzhōu . Fùzhōu -- . Xiāngtàn -- . M ěixiàn -- . H ongkong -- . ---Wǔhàn . ---Běijīng . ---Kùnmíng . Hángzhōu -- . Xiàmén -- . ---Chéngdū . Sùzhōu -- . Shànghǎi -- . Táiběi -- . ---Zhèngzhōu . Shèxiàn -- . ---Nánjīng . ---Guìyáng . W énzhōu -- . N ánníng -- . Tūnxī -- . ---Tiānjìn . Shāntóu -- . ---Xīníng . ---Q īngdǎo . ---Ürüm qi . ---Píngyáo . Nánchàng -- . ---Tàiyuán . Chángshā -- . Hǎikǒu -- . ---Héfèi . Jiàn'ǒu -- . ---Yīnchuàn . ---Hohhot . Táoyuán -- . ---Xī'ān . G uǎngzhōu -- . ---Harbin . ---Jìnán . 0 . 0 . 0 . Inferred Links 33 / 50

Slide 90

Slide 90 text

Slide 91

Slide 91 text

Slide 92

Slide 92 text

Slide 93

Slide 93 text

Quantitative Historical Linguistics Problems Problems 34 / 50

Slide 94

Slide 94 text

Quantitative Historical Linguistics Problems Transparency 35 / 50

Slide 95

Slide 95 text

Quantitative Historical Linguistics Problems Transparency Computational methods tend to be presented in a blackbox fashion. This does not mean that the methods are actually blackboxes upon which scholars do not have control, but rather that researchers avo- id the eﬀort of increasing the transparency of their algorithms in such a way that they log all relevant decisions that lead to a certain result. As a result, the majority of research published in quantitative historical linguistics is irreproducible and not falsiﬁable. 35 / 50

Slide 96

Slide 96 text

Quantitative Historical Linguistics Problems Applicability 36 / 50

Slide 97

Slide 97 text

Quantitative Historical Linguistics Problems Applicability Method Multilingual? No additional requirements? Freely Available? Mackay & Kondrak 2005 ✗ ✓ ✗ Bergsma & Kondrak 2007 ✓ ✓ ✗ Turchin et al. 2010 ✓ ✓ ✓ Berg-Kirkpatrick & Klein 2011 ✗ ✓ ✗ Hauer & Kondrak 2011 ✓ ✓ ✗ Steiner et al. 2011 ✓ ✓ ✗ List 2012 & List 2014 ✓ ✓ ✓ Beinborn et al. 2013 ✗ ? ✗ Bouchard-Côté et al. 2013 ✓ ✗ ✗ Rama 2013 ✗ ✓ ✗ Ciobanu & Dinu 2014 ✗ ✓ ✗ … … … … 36 / 50

Slide 98

Slide 98 text

Quantitative Historical Linguistics Problems Applicability Method Multilingual? No additional requirements? Freely Available? Mackay & Kondrak 2005 ✗ ✓ ✗ Bergsma & Kondrak 2007 ✓ ✓ ✗ Turchin et al. 2010 ✓ ✓ ✓ Berg-Kirkpatrick & Klein 2011 ✗ ✓ ✗ Hauer & Kondrak 2011 ✓ ✓ ✗ Steiner et al. 2011 ✓ ✓ ✗ List 2012 & 2014 ✓ ✓ ✓ Beinborn et al. 2013 ✗ ? ✗ Bouchard-Côté et al. 2013 ✓ ✗ ✗ Rama 2013 ✗ ✓ ✗ Ciobanu & Dinu 2014 ✗ ✓ ✗ … … … … 36 / 50

Slide 99

Slide 99 text

Quantitative Historical Linguistics Problems Applicability Method Multilingual? No additional requirements? Freely Available? Mackay & Kondrak 2005 ✗ ✓ ✗ Bergsma & Kondrak 2007 ✓ ✓ ✗ Turchin et al. 2010 ✓ ✓ ✓ Berg-Kirkpatrick & Klein 2011 ✗ ✓ ✗ Hauer & Kondrak 2011 ✓ ✓ ✗ Steiner et al. 2011 ✓ ✓ ✗ List 2012 & 2014 ✓ ✓ ✓ Beinborn et al. 2013 ✗ ? ✗ Bouchard-Côté et al. 2013 ✓ ✗ ✗ Rama 2013 ✗ ✓ ✗ Ciobanu & Dinu 2014 ✗ ✓ ✗ … … … … 36 / 50

Slide 100

Slide 100 text

Quantitative Historical Linguistics Problems Applicability Method Multilingual? No additional requirements? Freely Available? Mackay & Kondrak 2005 ✗ ✓ ✗ Bergsma & Kondrak 2007 ✓ ✓ ✗ Turchin et al. 2010 ✓ ✓ ✓ Berg-Kirkpatrick & Klein 2011 ✗ ✓ ✗ Hauer & Kondrak 2011 ✓ ✓ ✗ Steiner et al. 2011 ✓ ✓ ✗ List 2012 & 2014 ✓ ✓ ✓ Beinborn et al. 2013 ✗ ? ✗ Bouchard-Côté et al. 2013 ✓ ✗ ✗ Rama 2013 ✗ ✓ ✗ Ciobanu & Dinu 2014 ✗ ✓ ✗ … … … … 36 / 50

Slide 101

Slide 101 text

Quantitative Historical Linguistics Problems Accuracy 37 / 50

Slide 102

Slide 102 text

Slide 103

Slide 103 text

Quantitative Historical Linguistics Problems Accuracy Data Problems (Geisler & List forthcoming) Comparing two independently produced lexicostatistical datasets: database # languages # concepts Dyen et al. 1997 95 200 Tower of Babel 98 110 intersection 46 103 Results up to 10 % diﬀerence in concept translations many undetected borrowings in both datasets up to 30 % diﬀerences in tree topologies for Bayesian analyses 37 / 50

Slide 104

Slide 104 text

Quantitative Historical Linguistics Problems Accuracy 37 / 50

Slide 105

Slide 105 text

Quantitative Historical Linguistics Problems Comparability 38 / 50

Slide 106

Slide 106 text

Quantitative Historical Linguistics Problems Comparability The data which is used to test new quantitative methods varies greatly, and there seems to be little interest in standardization. As a result, it is diﬃcult if not impossible to compare the accuracy of dif- ferent methods in quantitative historical linguistics, and it drastically hampers the improvement of existing applications. 38 / 50

Slide 107

Slide 107 text

Quantitative Historical Linguistics Problems Comparability Data of an analysis of Turkic languages (Hruschka et al. 2015) 38 / 50

Slide 108

Slide 108 text

Quantitative Historical Linguistics Problems Summary Aspect Current State Desired State transparency blackbox methods prevail ALL results of should be presented in human- readable format applicability lack of free software applications with transpa- rent source code ALL software proposed should be free, and source code should be well commented adequacy adequacy of results is often questionable ALL methods need to be rigorously tested before publication comparability lack of standardization, inadequate biological formats prevail data representation and method testing needs to be standardized 39 / 50

Slide 109

Slide 109 text

Computer-Assisted Language Comparison 40 / 50

Slide 110

Slide 110 text

Computer-Assisted Language Comparison Bridging the Gap Bridging the Gap Computational approaches in historical linguistics largely disregard the actual needs of historical linguistics. Despite the frequent claims that the algorithms are intended to supple- ment traditional research, many of them are mere attempts to prove the power of modern machine learning approaches. On the other hand, traditional approaches often disregard the fascinating possibilities that the digital age oﬀers, and fail to leave the “realm of intuition”. 41 / 50

Slide 111

Slide 111 text

Computer-Assisted Language Comparison Bridging the Gap Bridging the Gap P(A|B)=(P(B|A)P(A))/(P(B) FRANZ BOPP VERY, VERY LONG TITLE 41 / 50

Slide 112

Slide 112 text

Computer-Assisted Language Comparison Bridging the Gap Bridging the Gap P(A|B)=(P(B|A)P(A))/(P(B) FRANZ BOPP VERY, VERY LONG TITLE 41 / 50

Slide 113

Slide 113 text

Computer-Assisted Language Comparison Bridging the Gap Bridging the Gap P(A|B)=(P(B|A)P(A))/(P(B) FRANZ BOPP VERY, VERY LONG TITLE Apart from “computational historical linguistics”, we need to establish a new discipline of “computer- assisted historical linguistics” (CALC). 41 / 50

Slide 114

Slide 114 text

Computer-Assisted Language Comparison Bridging the Gap Excursus: When do we need computers? Aspect Computers Needed? Example data Yes, if the 10 times rule of repeated action in data preparation applies! segmentation of IPA- encoded strings into meaningful sound units modeling Yes, since formalization serves as a proof of concept, guarantees comparability, and helps avoiding errors! family tree model in its current form inference Yes, since a workﬂow of computational analysis ac- companied by manual cor- rection is much faster! phonetic alignment and cognate detection analysis Yes, if the analysis invol- ves calculations or repeated actions! ﬁnding the tree that ex- plains a dataset best 42 / 50

Slide 115

Slide 115 text

Computer-Assisted Language Comparison Examples Examples: Data Preparation with help of the EDICTOR data preparation with help of the EDICTOR (List forthcoming, http://tsv.lingpy.org) data modeling (phonetic alignments, cognate judgments) with help of the EDICTOR inference of probable semantic shift patterns with help of CLICS (List et al. 2014, http://clics.lingpy.org) analysis of lexical change patterns with help of the MLN approach (List et al. 2014) 43 / 50

Slide 116

Slide 116 text

Computer-Assisted Language Comparison Challenges Challenges 44 / 50

Slide 117

Slide 117 text

Computer-Assisted Language Comparison Challenges Challenges German m oː n t - English m uː n - - Danish m ɔː n - ə Swedish m oː n - e 44 / 50

Slide 118

Slide 118 text

Computer-Assisted Language Comparison Challenges Challenges German m oː n t - English m uː n - - Danish m ɔː n - ə Swedish m oː n - e Fúzhōu ŋ u o ʔ ⁵ - - - - - - - - - - Měixiàn ŋ i a t ⁵ - - - - - k u o ŋ ⁴⁴ Guǎngzhōu j - y t ² l - œ ŋ ²² - - - - - Běijīng - y ɛ - ⁵¹ l i ɑ ŋ - - - - - - 44 / 50

Slide 119

Slide 119 text

Slide 120

Slide 120 text

Computer-Assisted Language Comparison Challenges Challenges Fúzhōu Měixiàn Guǎngzhōu Běijīng 44 / 50

Slide 121

Slide 121 text

Computer-Assisted Language Comparison Challenges Challenges Fúzhōu Měixiàn Guǎngzhōu Běijīng INNO VATIO N INNO VATIO N INNO VATIO N BO RRO W ING LO SS INNO VATIO N INNO VATIO N 44 / 50

Slide 122

Slide 122 text

Computer-Assisted Language Comparison Challenges Challenges SEMANTIC CHANGE MORPHOLOGICAL CHANGE S T R A T IC C H A N G E Three Dimensions of Lexical Change (Gévaudan 2007) 44 / 50

Slide 123

Slide 123 text

魚 1 魚 1 魚 1 ? 首首首首 ... and its application to Chinese dialect History 45 / 50

Slide 124

Slide 124 text

Application to Chinese Dialect History DFG Research Project DFG Research Project on Chinese Dialect History My research project “Vertical and lateral aspects of Chinese dialect history”, funded by the German Research Foundation (DFG), pro- poses an interdisciplinary, data-driven approach to Chinese dialectology and has two major goals: 1 Based on the cooperation with sinologists (CLRAO) and biologists (UMPC), quantitative methods originally designed to study lateral gene transfer in evolutionary biology shall be used to explore vertical (inheritance-related) and lateral (contact-related) aspects of Chinese dialect history. 2 In order to bridge the gap between traditional and quantitative approaches in historical linguistics the research will be carried out within the new framework of “computer-assisted language comparison”. 46 / 50

Slide 125

Slide 125 text

Application to Chinese Dialect History Data Preparation Data Preparation Hànyǔ Fāngyán Cíhuì 汉语方言词汇 (Běijīng Dàxué 1964, 904 concepts, translated into 17 dialect varieties) serves as the fundamental data basis, hopefully expanded by further sources provided by colleagues or taken from additional literature digital version of the Cíhuì has already been prepared about 80 percent of all Chinese glosses have already been translated to Chinese during the next two weeks, all data will be systematically extracted, cleaned (segmentation of phonetic entries, e.g. write "thai51jaŋ35" as "tʰ a j ⁵¹ j a ŋ ³⁵"), and uploaded to the EDICTOR tool for further analysis (as part of a larger database on Sino-Tibetan languages) additional sources will be subsequently added (web-based tools for quick identiﬁcation of identical concepts in new sources have already been written) 47 / 50

Slide 126

Slide 126 text

Application to Chinese Dialect History Modeling Modeling all lexical entries will be searched for cognate parts cognate sets will be aligned by ﬁrst carrying out automatic analyses and correcting them later manually each lexical entry will further be aligned with its Middle Chinese and Old Chinese ancestral forms => In this way, I hope to make sure that a maximal amount of knowledge about Chinese dialect history is represented in a human- and machine-readable way and can later be accessed and analysed by both humans and computers. 48 / 50

Slide 127

Slide 127 text

Application to Chinese Dialect History Inference and Analysis Inference and Analysis In Chinese dialectology, it is often not very hard to detect whether two or more words are etymologically related. What is hard, however, is to determine their exact relationship, that is, it is hard to resolve which dimensions of lexical change were involved during their history. In order to cope with the multiple dimensions of lexical change, we need new methods and models in historical linguistics, which explicitly deal with borrowing, partial cognacy, and semantic change. Following the lead of evolutionary biology, these methods could be combined under a uniﬁed framework of tree reconciliation (Page & Cotton 2002) in historical linguistics. 49 / 50

Slide 128

Slide 128 text

Application to Chinese Dialect History Inference and Analysis Inference and Analysis Fúzhōu Měixiàn Guǎngzhōu Běijīng Fúzhōu Měixiàn Guǎngzhōu Běijīng 49 / 50

Slide 129

Slide 129 text

Application to Chinese Dialect History Inference and Analysis Inference and Analysis Fúzhōu Měixiàn Guǎngzhōu Běijīng Fúzhōu Měixiàn Guǎngzhōu Běijīng 49 / 50

Slide 130

Slide 130 text

Application to Chinese Dialect History Inference and Analysis Inference and Analysis Fúzhōu Měixiàn Guǎngzhōu Běijīng 49 / 50

Slide 131

Slide 131 text

Application to Chinese Dialect History Inference and Analysis Inference and Analysis Fúzhōu Měixiàn Guǎngzhōu Běijīng 49 / 50

Slide 132

Slide 132 text

Application to Chinese Dialect History Inference and Analysis Inference and Analysis LOSS INNO VATIO N INNO VATIO N BORROWING 49 / 50

Slide 133

Slide 133 text

Application to Chinese Dialect History Inference and Analysis Inference and Analysis LOSS INNO VATIO N INNO VATIO N BORROWING Tree-conciliation is but one method that can be used to provide ﬁrst heu- ristics to disentangle the complicated etymolo- gical relations between Chinese dialect words. Alternative methods will be discussed and tested with the biological cooperators during my stay. 49 / 50

Slide 134

Slide 134 text

In the end of the 19th century, many dialectologists began to oppose the Neogrammarian claim that sound change proceeds without exceptions. Under the slogan “chaque mot a son histoire”, they de- manded that linguists should take all available data, especially dialect data, into account when propo- sing their theories on language variation and change. This triggered some debates in historical linguistics, but they were soon forgotten and the methods were never changed. 50 / 50

Slide 135

Slide 135 text

Not long ago, evolutionary biologists realized that the evolution of genes may follow very idiosyncratic patterns, so complex indeed, that one could likewi- se claim that each gene has its own history. In order to account for this, they started to develop computational tools that allow them to study what they now sometimes call “mosaic history”, that is, the individual histories of genes that do not follow a major evolutionary path. 50 / 50

Slide 136

Slide 136 text

In the digital age we are confronted with an abun- dance of language data that no one person can re- ad in a lifetime. It seems to be time that historical linguists start following the example of evolutionary biology in using computers as a helpful tool to investigate our data and not giving easily up if our theories are threatened by reality. If it is really true, that every word has its own history, then let’s try to assemble all these histories in databases and look which new and interesting patterns we may ﬁnd! 50 / 50

Slide 137

Slide 137 text

Merci pour votre attention! 50 / 50