The overarching issue explored in the project is cross-linguistic variability in semantic systems and its limits in space and time. This was approached by focusing on the temperature domain and systematically comparing the systems of central temperature terms (hot, cold etc.) in 70 genetically and areally diverse languages from three angles: to what extent linguistic categorization is universal or language- and culture-specific, how semantic systems emerge and develop, and how expressions for categories in one semantic domain are used for others. The project is a sequel to Koptjevskaja-Tamm ed. 2015 (https://benjamins.com/catalog/tsl.107), which contained detailed descriptions of the temperature systems in 50 languages, but no systematic comparison apart from several tentative generalizations.

A part of the project time was spent on two stays at the research centers LLACAN / LACITO (CNRS) in Paris. They were devoted to the (re-)analysis, systematization and (re-)interpretation of the earlier data and complementing them with data on the same or new languages, in intensive cooperation with the French specialists in African and Austronesian languages and cultures. The initial plan to complete a monograph turned out not to be feasible within the allocated period and was therefore replaced by a combination of the time-consuming work on the monograph with the work on several journal articles and chapters (often co-authored) devoted to particular aspects of the overarching project. These are presented below.

1. A lexico-constructional typology of temperature expressions
Traditionally, lexical typology has focused on how cognitive domains e.g. COLOUR, are cut up by words in a language, while their grammatical behavior and morphosyntactic patterns traditionally belong to grammatical typology. Cross-linguistic generalizations are often represented by semantic maps, normally restricted to one particular kind of meaning – lexical, constructional or grammatical – and there is no obvious method to represent the interface between lexicon and grammar, which should be at the core of lexical typology. Importantly, the tentative partial generalizations in Koptjevskaja-Tamm ed. 2015 apply either to purely lexical (e.g., separate words for ‘hot’ in ‘The tea is hot’, ‘It is hot here’ and ‘I feel hot’) or to purely grammatical phenomena (e.g., the different constructions corresponding to these differences), even though the encoding of temperature involves their interaction. Major efforts in the project were therefore directed towards the theoretical and methodological foundations of lexical typology, the analysis of the new data and the re-analysis and re-conceptualisation of the old data. Koptjevskaja-Tamm & Vejdemo (2019) argue that categories emerging in the course of cross-linguistic research seem to be organized in terms of prototypes, but there is relatively little discussion of prototype semantics and focal meanings in lexical typology. Koptjevskaja-Tamm & Veselinova’s (2019) programmatic chapter gives a broad outline of the areas where lexical typology and morphology share interests often without being aware of this connection, thereby calling for a more direct interaction between the two. Koptjevskaja-Tamm (in prep.) is integrating semantic maps with Construction-Grammar inspired schemas for representing the lexico-constructional typology of temperature expressions.

2. Evolution of temperature systems
There is a long tradition of research on lexical/semantic change, etymological sources and lexical replacement, recently boosted by the use of advanced statistical modelling. However, these studies normally focus on individual words, whereas closely related languages often show significant typological differences in categorization of whole cognitive domains. The comparison of the temperature adjectives across Slavic in Koptjevskaja-Tamm (2019) is ground-breaking in targeting both the systems as a whole and the forms involved in them. The Slavic comparison serves as the model for the current work on Germanic (in cooperation with Thomas de Roo, a MA student at the University of Groningen), Sinitic (in cooperation with Junfan Lai, a postdoc researcher at the Max-Planck Institute for the Science of Human History at Jena), Romance and Uralic.

3. Extended uses of temperature expressions
This part of the project targets uses of temperature expressions outside of the temperature domain, including metaphors such as warm words, investigates the areal and genetic patterns in their distribution and asks to what extent the findings provide evidence for vs. against the allegedly universal conceptual metaphors and semantic shifts in general. Major efforts were made to systematize the old data and to complement them with new. The research confirmed the earlier observations that there are clear areal differences in the extended uses of temperature terms. The link between warmth and affection in temperature words is e.g. primarily (though not exclusively) found in the languages of Europe, while the African languages often have extended uses of ‘cold’ with positive associations, such as ‘peace’ or ‘calm’. The connection between heat and intensity, including anger and danger, is widely attested all over the world, even though many languages completely lack extended uses of temperature words. The two major sources of data on extended uses are the lexical databases CLICS2 (https://clics.clld.org) and RefLex (https://sites.google.com/site/referencelexicon/), both invaluable resources but each one with its limitations. The work with the databases has led to three joint publications focusing on several important methodological and theoretical issues. Gast & Koptjevskaja-Tamm (2018, subm.) suggest, test and evaluate “quick and dirty” methods for automatically extracting areal, genetic and “more universal” colexification patterns from lexical databases. Rzymski et al. (subm.) is an innovative “spin-off” of Gast & Koptjevskaja-Tamm (2018, subm.) and of the whole project, based on the experience of working with CLICS2. Koptjevskaja-Tamm et al. (in prep.) is intended to be a ground-breaking contribution to the emergent and under-explored field of areal lexico-semantics, i.e. the diffusion of semantic features across language boundaries, especially in one geographic area. The findings related to the ‘affection as warmth’ expressions and their physiological and cultural underpinning have been discussed with social psychologists and cultural anthropologists. A brief overview of the concrete and extended uses of temperature expressions in the world’s languages intended for researchers in social thermoregulation will appear in Koptjevskaja-Tamm (in prep. a).

The project’s three most important results:

1. An integrated lexico-constructional typology of temperature expressions as a contribution to the issue of semiotic ecology, i.e. how semantic choices made in one linguistic subsystem (lexicon) affect those in others (grammar)

2. A new approach to a cross-linguistic comparison of the lexical systems in closely-related languages targeting both individual words and systems on the whole

3. Areal patterns in the distribution of temperature-based metaphors, including the limited attestation of ‘affection as warmth’ uses of the temperature expressions

New research questions generated through the project.

1. The temperature systems across languages are often heterogeneous in that their different subparts behave differently. Most languages also lack the abstract concept of temperature. Is the temperature domain a linguistic construct or are there arguments for its independent cognitive status? And, more generally, how do we recognize cognitive domains?

2. Do semantic patterns differ in how easily they can diffuse via language contact? If yes, why?

3. Lexical typology is on the rise, but the studies on the different domains are seldom related to each other. What overarching generalizations stem from lexical-typological research?

The project’s main international collaboration partners

The social psychologist Hans IJzerman, Université Grenoble Alpes, and the Kama Muta Lab at the University of Oslo (http://kamamutalab.org), primarily the social psychologists Beate Seibt and Thomas Schubert (Oslo), and the psychological anthropologist Alan P. Fiske (UCLA) (social thermoregulation)

LLACAN (Languages and cultures of Sub-Saharan Africa, http://llacan.vjf.cnrs.fr/index_en.php) and LACITO (Langues et Civilisations à Tradition Orale http://lacito.vjf.cnrs.fr/themes/index.htm) within the French National Centre for Scientific Research (CNRS)

Research Unit “Mondes Anciens” at the University of Liège (semantic maps)
Moscow Lexical Typology (the National Research University High School of Economics, Moscow)

Volker Gast (the University at Jena) and Mattis List (the Max-Planck Institute for the Science of Human History at Jena) (work on lexical databases)

Matti Miestamo, Helsinki University (antonymy)

The work on the project and its results were presented in several seminars at CNRS in Paris and at the Kama Muta Lab in Oslo. MKT was an invited speaker at the conference on semantic maps at the University at Liège in June 2019, and a plenary speaker at the Societas Linguistica Europaea (SLE) annual conference in Tallinn in August 2018. MKT’s PhD course on semantics in contact at the SLE summer school in Tartu (August 2018) featured many aspects of the work in the project. Turning to the broader audience, MKT has been interviewed by the linguist and radio journalist David Midgley for the radio program "Talk the Talk" in Perth, Australia (broadcasted in February 2019) (http://talkthetalkpodcast.com/313-thats-hot-thats-cool/), and has given a lecture on the linguistics of temperature at the Franska skolan in Stockholm.

MKT’s stays at LLACAN/LACITO were in addition used for planning a PhD summer school on semantic shifts that will be organized by CNRS with the Department of Linguistics, Stockholm University as co-organizer on the island of Porquerolles in September 2020.