���Keep CALM and carry on doing the comparative method �(just not scientistically)��Christopher Lucas (SOAS University of London)�Lameen Souag – LACITO (CNRS – Sorbonne Nouvelle – INALCO) – NILOMORPH��

ICHL 27, Santiago, 18 August 2025

Funded by the European Union. Views and opinions expressed �are however those of the author(s) only and do not necessarily �reflect those of the European Union or the European Research �Council Executive Agency. Neither the European Union nor the �granting authority can be held responsible for them.

​

The discipline of linguistics

2

• Linguistics (esp. historical linguistics) straddles humanities and the sciences.
• One way of understanding the distinction:
• Humanities: sharpens our understanding of what we value (diversity, cultural heritage, etc.)
• Science: seeks to make discoveries about the natural world (including human beings)
• The present talk attempts to sharpen our understanding of the value (and limits) of the comparative method (i.e. we’re doing humanities).

​

Scientific discoveries

3

• Discoveries: initially counterintuitive, but demonstrably correct and ~universally accepted, analyses of nature.
• E.g.: heliocentrism, plate tectonics, “ballooning” in spiders by means of electrostatic repulsion (Morley & Robert 2018), and…
• The comparative method (CM). In particular:
• Exceptionlessness: sounds in a variety evolve identically in a given phonological context across all inherited lexical items.
• Relatedness is thus demonstrated through systematic sound correspondences, not similarity.

​

What makes CM different?

4

• How to understand the enduring success (and limitations) of CM in different contexts…
• …contrasting with the lack of discoveries in most of theoretical linguistics (where instead we just have temporarily popular trends)?
• McGinn’s (1993) epistemology of CALM: combinatorial atomism with lawlike mappings

​

​

Understanding CALM

5

• Human cognitive capacity, like any other species, is naturally evolved and specialized.
• Not reasonable to think we are equally well equipped to understand all facets of the universe.
• Where we excel:
• Systems with CALM features (e.g. inorganic chemistry)
• We find the “atoms” of a system and establish the laws governing how they interact to form complex wholes.

​

Understanding CALM

6

• Where we make no progress and establish no consensus:
• Systems lacking CALM features, no obvious “atoms”.
• Paradigmatic example: consciousness
• But also many areas of the hard sciences, e.g.: what exactly are time and space?

​

​

​

Smolin (2006)

Sound change is CALM

7

• Much of language is not CALM: hence many linguistic theories but no consensus.
• Sound change is overwhelmingly CALM:
• Atoms are speech segments (phones)
• Further decomposable into features of place, manner, voicing etc.
• Interaction of features and segments in lexical items is highly regular: lawlike mapping.
• Hence consensus, replicability, progress.

​

Reductiveness

8

• Key feature of CM and all CALM science: it is reductive.
• Sound change is only exceptionless once we abstract away from loanwords, analogical formations, etc.
• CM only tells us whether/how “languages” are related (in fact only parts of lexicon + phonology).
• Without reduction: no discoveries, only debates.
• With reduction: discoveries about small slices of nature (the parts that are CALM).

Reductiveness vs scientism

9

• Responsible reductiveness is thus crucial: should not be a pejorative term!
• Irresponsible reductiveness is called scientism (e.g. Haack 2003):
• Wrongly sees what is abstracted away from as irrelevant or illusory
• Dismisses approaches that do not use CALM reasoning
• We should defend CM but understand that it is just one valuable tool among many in historical linguistics.

​

Case studies from Arabic: 1

10

• Classical Arabic (CA) is amply attested and highly archaic.
• But can we simply assume it is equivalent to the ancestor of all present-day spoken varieties?
• Owens (2005) notes a feature, absent from CA, found only in a few widely dispersed dialects spoken on the periphery of the Arabic-speaking world:
• A semantically empty -in(n)- affix that obligatorily appears between an active (resultative) participle and an object pronoun suffix:

Case studies from Arabic: 1

11

Case studies from Arabic: 1

12

• Since this feature is so arbitrary, we must be dealing with common inheritance, not independent parallel developments.
• The common ancestor of these three dialects, which predates attested CA, must be reconstructed as having this morpheme.
• We could not know this about early Arabic without the CM.
• We only discovered it thanks to Owens’ (2005) reconstruction.

​

Case studies from Arabic: 2

13

• A cautionary tale: Jaradat (2024) offers a purely internal reconstruction of the prehistory of the demonstrative element ðaː.
• He suggests, counter to typical grammaticalization paths, that it derives from ðaːt ‘self’, via ðuː/ðaː/ðiː ‘the one in possession of’.
• But there is no reason not to use the CM here, which shows that the demonstrative function for ðaː goes back at least to Proto-West-Semitic. Jaradat’s internal reconstruction is therefore untenable.

​

​

​

Case studies from Arabic: 2

14

• Comparative West Semitic data, with ‘male’ showing the regular correspondences of Arabic ð in other languages:

​

​

Hasselbach (2007)

Case studies from Arabic: 3

15

• Syntactic reconstruction is, at best, more tentative than orthodox lexical reconstruction via the CM, due to the lack of “double cognacy” (Walkden 2013): sounds in lexical items are cognate, as are the items themselves.
• Caution is therefore required when reconstructing the pre-history of apparently parallel syntactic structures in related varieties.
• Owens (2023: 227–9) and Wilmsen (2025) note instances of “clitic doubling” in disparate Arabic varieties (Levantine, Iraqi, Maltese, Maghrebi, among others) and assume that all are reflexes of common inheritance.

​

​

Case studies from Arabic: 3

16

• But Souag (2017) shows that Arabic varieties match their non-Arabic neighbours better than they match each other

​

​

Case studies from Arabic: 3

17

• The initial impression of parallelism results from a holistic conception of “clitic doubling”, lumping together multiple constructions based on a single prominent trait
• (the presence of a clitic pronoun and a co-referential noun phrase expressing the same argument)
• Stay CALM: combinatorial atomism allows a better historical analysis here
• But the CM by itself does not – it neglects contact effects

​

​

Conclusion

18

• The CM is a landmark discovery of CALM linguistics, which remains as useful and relevant—worldwide—as it was 150 years ago.
• But its scope is limited! It should be used always and only where appropriate.
• It is one useful tool among many in diachronic linguistics.
• We should preserve and celebrate it, but not engage in scientism: other, non-CALM tools and approaches are equally valuable for a comprehensive understanding of linguistic history.

​

​

​

19

​

​

Thank you!

​

20

Haack, Susan. 2003. Defending science – within reason: Between scientism and cynicism. Amherst, NY: Prometheus Books.

Hasselbach, Rachel, 2007. Demonstratives in Semitic. Journal of the American Oriental Society, 127(1). 1–27.

Jaradat, Abdulazeez. 2024. From nominal source to demonstrative: A case of grammaticalization in Standard Arabic. STUF - Language Typology and Universals 77(3). 315–352.

Morley, Erica L & Daniel Robert. 2018. Electric fields elicit ballooning in spiders. Current Biology. 28(14). 2324–2330.

Owens, Jonathan. 2005. Pre-diaspora Arabic: Dialects, statistics and historical reconstruction. Diachronica 22. 271–308.

Owens, Jonathan. 2023. Arabic and the case against linearity in historical linguistics. Oxford: Oxford University Press.

Smolin, Lee. 01/02/2006. Atoms of space and time. Scientific American.

Souag, Lameen. 2017. Clitic doubling and contact in Arabic. Zeitschrift für Arabische Linguistik 66. 45–70.

Walkden, George. 2013. The correspondence problem in syntactic reconstruction. Diachronica 30(1). 95–122.

Wilmsen, David. 2025. The origin of object reduplication in Maltese: Two contacts, one, or none? In Rafaello Bezzina, Michela Vella & Maike Vorholt (eds.), The next century of Maltese linguistics. Berlin: De Gruyter Mouton.

​

​

​

​

​

​

References