Biology and Linguistics Johann-Mattis List DFG research fellow Centre des recherches linguistiques sur l’Asie Orientale Team Adaptation, Integration, Reticulation, Evolution EHESS and UPMC, Paris 2016/05/09 1 / 47
Schleicher (1821-1868) “These assumptions, which follow logically from the results of our re- search, can be best illustrated by the image of a branching tree.” (Schle- icher 1853: 787) 3 / 47
Schmidt (1843-1901) “You can turn it as you want, but as long as you stick to the idea that the his- torically attested languages have been developing by multiple furcations of an ancestral language, that is, as long as you assume that there is a Stammbaum [family tree] of the Indo-European lan- guages, you will never be able to explain all facts which have been assembled in a scientifically satisfying way.” (Schmidt 1872: 17, my translation) 5 / 47
Schmidt (1843-1901) “I want to replace [the tree] by the im- age of a wave that spreads out from the center in concentric circles be- coming weaker and weaker the far- ther they get away from the center.” (Schmidt 1872: 27, my translation) 6 / 47
Trees are bad, because... they are difficult to reconstruct............ languages do not always split............ .......... ............ ............ 8 / 47
Trees are bad, because... they are difficult to reconstruct............ languages do not always split............ .......... ............ ............ they are boring, since they only model the vertical aspects of language history ............ 8 / 47
Trees are bad, because... they are difficult to reconstruct............ languages do not always split............ .......... ............ ............ they are boring, since they only model the vertical aspects of language history ............ Waves are bad, because nobody knows how to reconstruct them 8 / 47
Trees are bad, because... they are difficult to reconstruct............ languages do not always split............ .......... ............ ............ they are boring, since they only model the vertical aspects of language history ............ Waves are bad, because nobody knows how to reconstruct them languages still diverge, even if not necessarily in split processes 8 / 47
Trees are bad, because... they are difficult to reconstruct............ languages do not always split............ .......... ............ ............ they are boring, since they only model the vertical aspects of language history ............ Waves are bad, because nobody knows how to reconstruct them languages still diverge, even if not necessarily in split processes they are boring, since they only model the horizontal aspects of language history 8 / 47
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) 9 / 47
Indo-European soːwel- sunːoː- SUN Germanic soːl- SUN soːlikul- SMALL SUN Romance solej SUN French sol SUN Spanish zɔnə SUN German suːl SUN Swedish SEM ANTIC SHIFT M O RPH O LO G ICAL CH AN G E M O R PH O LO G ICA L CH A N G E MORPHOLOGICAL CHANGE MORPHOLOGICAL CHANGE 11 / 47
pluːma pjuma ‘FLAT’ plaːnus pjano ‘SQUARE’ plateːa pjaʦːa Meaning Latin Italian ‘TONGUE’ liŋgua liŋgwa ‘MOON’ lu:na luna ‘SLOW’ lentus lento l > j 13 / 47
pluːma pjuma ‘FLAT’ plaːnus pjano ‘SQUARE’ plateːa pjaʦːa Meaning Latin Italian ‘TONGUE’ liŋgua liŋgwa ‘MOON’ lu:na luna ‘SLOW’ lentus lento l > j l > l 13 / 47
pluːma pjuma ‘FLAT’ plaːnus pjano ‘SQUARE’ plateːa pjaʦːa Meaning Latin Italian ‘TONGUE’ liŋgua liŋgwa ‘MOON’ lu:na luna ‘SLOW’ lentus lento l > j l > l l > j / p _ 13 / 47
pluːma pjuma ‘FLAT’ plaːnus pjano ‘SQUARE’ plateːa pjaʦːa Meaning Latin Italian ‘TONGUE’ liŋgua liŋgwa ‘MOON’ lu:na luna ‘SLOW’ lentus lento l > j l > l l > j / p _ Not sounds change, sound systems change (Bloomfield 1933)! 13 / 47
pluːma pjuma ‘FLAT’ plaːnus pjano ‘SQUARE’ plateːa pjaʦːa Meaning Latin Italian ‘TONGUE’ liŋgua liŋgwa ‘MOON’ lu:na luna ‘SLOW’ lentus lento l > j l > l l > j / p _ Not sounds change, sound systems change (Bloomfield 1933)! Sound change depends on the context in which the sounds occur! 13 / 47
pluːma pjuma ‘FLAT’ plaːnus pjano ‘SQUARE’ plateːa pjaʦːa Meaning Latin Italian ‘TONGUE’ liŋgua liŋgwa ‘MOON’ lu:na luna ‘SLOW’ lentus lento l > j l > l l > j / p _ Not sounds change, sound systems change (Bloomfield 1933)! Sound change depends on the context in which the sounds occur! Sound change largely follows irreversible patterns! 13 / 47
to the Past The Geological Evidences of The Antiquity of Man with Remarks on Theories of The Origin of Species by Variation By Sir Charles Lyell London John Murray, Albemarle Street 1863 15 / 47
to the Past If we new not- hing of the existence of Latin, - if all historical documents previous to the fin- teenth century had been lost, - if tra- dition even was si- lent as to the former existance of a Ro- man empire, a me- re comparison of the Italian, Spanish, Portuguese, French, Wallachian, and Rhaetian dialects would enable us to say that at some time there must ha- ve been a language, from which these six modern dialects derive their origin in common. 15 / 47
to the Past: Uniformitarianism (C. Lyell) Uniformity of Change: Laws of change are uniform. They have applied in the past as they apply now and will apply in the future, no matter at which place. 16 / 47
to the Past: Uniformitarianism (C. Lyell) Uniformity of Change: Laws of change are uniform. They have applied in the past as they apply now and will apply in the future, no matter at which place. Graduality of Change: Change proceeds gradually, not abrupt. 16 / 47
to the Past: Uniformitarianism (C. Lyell) Uniformity of Change: Laws of change are uniform. They have applied in the past as they apply now and will apply in the future, no matter at which place. Graduality of Change: Change proceeds gradually, not abrupt. Abductive Reasoning: We can infer past events and processes by investigating patterns observed in the present, which becomes the “key to the interpretation of some mystery in the archives of remote ages” (Lyell 1830: 165) 16 / 47
to the Past: Uniformitarianism (A. Schleicher) Language Change is a gradual process (Schleicher 1848: 25). is a law-like process (Schleicher 1848: 25). is a natural process which occurs in all languages (Schleicher 1848: 25). universal process which occurs in all times (Schleicher 1863[1873]: 10f). allows us to infer past processes and extinct languages by investigating the languages of the present (see Schleicher 1848: 25). 17 / 47
to the Past: Summary It was not the direct exchange of ideas that lead to the de- velopment of similar approaches in biology and linguistics, but the astonishing fact that scholars in both fields would at about the same time detect striking parallels between both disciplines, both regarding their theoretical founda- tions and the processes they were investigating. 18 / 47
to the Past: Summary It was not the direct exchange of ideas that lead to the de- velopment of similar approaches in biology and linguistics, but the astonishing fact that scholars in both fields would at about the same time detect striking parallels between both disciplines, both regarding their theoretical founda- tions and the processes they were investigating. And linguists were the first to draw trees! 18 / 47
to the Past: Summary 1700 1800 1750 1850 List et al. (in preparation) Stiernhielm's Lingua Nova 1671 Gallet's Arbre ca. 1800 Darwin's Origins 1859 De Buffon's Table 1755 Schleicher's Stammbaum 1853 Darwin's Tree Sketch 1837 Lamarck's Tableaux 1809 Čelakovský's Rodový Kmen 1853 Rühling's Tabula 1774 Hicke's Affinitas 1689 Schottels's Tabelle 1663 19 / 47
(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) “Indo-European languages tree by Levenshtein distance” (Serva and Petroni 2008) “Networks uncover hidden lexical borrowing in Indo-European language evolution” (Nelson-Sathi et al. 2011) 21 / 47
(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) “Indo-European languages tree by Levenshtein distance” (Serva and Petroni 2008) “Networks uncover hidden lexical borrowing in Indo-European language evolution” (Nelson-Sathi et al. 2011) 21 / 47
Concept German ID English ID Italian ID French ID HAND Hand 1 hand 1 mano 2 main 2 BLOOD Blut 3 blood 3 sangue 4 sang 4 HEAD Kopf 5 head 6 testa 7 tête 7 TOOTH Zahn 8 tooth 8 dente 8 dent 8 TO SLEEP schlafen 9 sleep 9 dormir 10 dormir 10 TO SAY sagen 11 say 11 dire 12 dire 12 ... ... ... ... ... ... ... ... ... 22 / 47
Concept German ID English ID Italian ID French ID HAND Hand 1 hand 1 mano 2 main 2 BLOOD Blut 3 blood 3 sangue 4 sang 4 HEAD Kopf 5 head 6 testa 7 tête 7 TOOTH Zahn 8 tooth 8 dente 8 dent 8 TO SLEEP schlafen 9 sleep 9 dormir 10 dormir 10 TO SAY sagen 11 say 11 dire 12 dire 12 ... ... ... ... ... ... ... ... ... 22 / 47
Concept German English Italian French “HAND” G E I F Hand 0 1 2 3 hand 1 0 2 3 mano 2 2 0 2 main 3 3 2 0 “BLOOD” G E I F Blut 0 4 5 4 blood 4 0 6 5 sangue 5 6 0 2 sang 4 5 2 0 Edit Distances between Orthographic Entries 23 / 47
Languages (Pagel 2009) aspect species languages unit of replication gene word replication asexual und sexual reproduction learning speciation cladogenesis language split forces of change natural selection and genetic drift social selection and trends differentiation tree-like tree-like 24 / 47
Languages (Geisler & List 2013) Aspect Species Languages domain Popper’s World I Popper’s World III relation between form and function mechanical arbitrary origin monogenesis unclear sequence similarity universal (indepen- dent of species) language-specific differentiation tree-like network-like 26 / 47
coined by Richard Owen (1804–1892), who distinguished ‘homologues’, as ‘the same organ in different animals under every variety of form and function’ (Owen 1843: 379), from from ‘analogues’ as an ‘organ in one animal which has the same function as another part or organ in a different animal’ (ibid.: 374). 29 / 47
coined by Richard Owen (1804–1892), who distinguished ‘homologues’, as ‘the same organ in different animals under every variety of form and function’ (Owen 1843: 379), from from ‘analogues’ as an ‘organ in one animal which has the same function as another part or organ in a different animal’ (ibid.: 374). Nowadays, it commonly denotes a ‘relationship of common descent between any entities, without further specification of the evolutionary scenario’ (Koonin 2005: 311). 29 / 47
coined by Richard Owen (1804–1892), who distinguished ‘homologues’, as ‘the same organ in different animals under every variety of form and function’ (Owen 1843: 379), from from ‘analogues’ as an ‘organ in one animal which has the same function as another part or organ in a different animal’ (ibid.: 374). Nowadays, it commonly denotes a ‘relationship of common descent between any entities, without further specification of the evolutionary scenario’ (Koonin 2005: 311). With respect to specific scenarios of common descent, molecular biologists characterize relationships between homologous genes further by distinguishing between orthology, paralogy, and xenology. 29 / 47
Indo-European *deh₃- *deh₃-no- Latin dare dōnum dōnāre Italian sole French soleil Swedish sol German Sonne Germanic *sōwel- *sunnō- Latin sol soliculus Indo-European *sóh₂-wl̩ - *sh₂én- A B List (in press) 30 / 47
frequent Latin words (Stefenelli 1992), 67% were directly inherited in at least one of the descendant languages of Latin, 14% were directly inherited in all descendant languages, 32 / 47
frequent Latin words (Stefenelli 1992), 67% were directly inherited in at least one of the descendant languages of Latin, 14% were directly inherited in all descendant languages, only 33% are completely lost, 32 / 47
frequent Latin words (Stefenelli 1992), 67% were directly inherited in at least one of the descendant languages of Latin, 14% were directly inherited in all descendant languages, only 33% are completely lost, about 50% of the words survive as borrowings from Latin in the descendant languages 32 / 47
frequent Latin words (Stefenelli 1992), 67% were directly inherited in at least one of the descendant languages of Latin, 14% were directly inherited in all descendant languages, only 33% are completely lost, about 50% of the words survive as borrowings from Latin in the descendant languages Saying that languages evolve in tree-like processes is similar to saying that penguins walk: It may be true, but it’s only a part of the whole interesting story. 32 / 47
we sequence 61 human genomes, we will find more or less the same collection of about 30,000 genes in each individual. But if we sequence 61 genomes of Escherichia coli (Lukjancenko et al. 2010) we find about 4,500 genes in each individual, we find 1,000 genes present in all genomes, we find about 18,000 different genes distributed among all genomes. 34 / 47
and Prokaryotic Evolution Eukaryotic populations generate tree-like divergence structures over time, while genome evolution in prokaryotes generates both tree-like and net-like components. Evolution and Language History Recalling the scores on borrowing frequency in the descendant languages of Latin, it seems obvious that language history shows a much closer resemblance to prokaryotic evolution than to eukaryotic evolution. When trying to apply methods from bioinformatics to linguistic problems, it seems therefore more fruitful to use those methods that explicitly deal with prokaryotic evolution. 35 / 47
Networks Biological Workflow (Dagan and Martin 2007, Dagan et al. 2008) 1 collect phyletic pattern data (shared gene families) of the taxa that shall be investigated 2 use gain-loss mapping techniques with different weighting models, allowing for different amounts of gain events to analyze how the gene families evolved along a given reference tree 3 use ancestral genome sizes as an external criterion to determine the best weighting model 4 assume that all patterns for which the best model yields more than one gain event result from lateral gene transfer 5 reconstruct a minimal lateral network by connecting multiple gains for the same gene family by lateral edges 36 / 47
Networks Linguistic Workflow (Nelson-Sathi et al. 2011, List et al. 2014) 1 collect phyletic pattern data (shared cognates) of the languages that shall be investigated 2 use gain-loss mapping techniques with different weighting models, allowing for different amounts of to analyze how the cognates evolved along a given reference tree 3 use ancestral vocabulary size distributions as an external criterion to determine the best weighting model 4 allow for a substantial amount (5%) of parallel evolution 5 assume that all patterns for which the best model yields more than one gain event result from lateral gene transfer 6 reconstruct a minimal lateral network by connecting multiple gains of the same cognate by lateral edges 37 / 47
Data (List et al. 2014a) Data 40 Indo-European languages (taken from the IELex, Dunn 2012) 1190 cognate sets (207 semantic glosses) 105 cognate sets contain known borrowings traditional reference tree, reflecting a very broad consensus, taken from Ethnologue (Lewis and Fennig 2013) 39 / 47
Data (List et al. 2014a) Analysis bottom-up parsimony-based approach for gain-loss mapping using different weight ratios for gain and loss events modified analysis allows for multifurcating (polytomic) reference trees specific factor for parallel evolution was added to the evaluation procedure implementation as part of the LingPy Python library for quantitative tasks in historical linguistics (http://lingpy.org, Version 2.2, List et al. 2013) 39 / 47
Data (List et al. 2014a) Results 76 cognate sets correctly identified as borrowings 31% of all cognate sets could not be properly explained by the reference tree 17 out of 19 borrowings in English correctly identified well-known contact situations among major groups and languages were correctly identified 39 / 47
in historical linguistics, we need to increase the complexity of our models. In order to profit from existing biological approaches, we need to foster a close collaboration between computational linguistics, computational biologists, but also and especially between classical biologists and linguistics. We need to be extremely careful with our analogies. Instead of transfer methods blindly, we should not forget that we may have to adapt them first to meet the specific needs of the target discipline. 43 / 47
lingulist
dropout009
norihirosuzuki
gully
keyakkie
axjack
tok41
ynakano
ngehlenborg
bob3bob3
nyoho
paulrobertlloyd
holman
dougneiner
malarkey
jacobian
trishagee
sstephenson
chriscoyier
schacon
colly
eitanlees
eileencodes
