LINGUIST List 14.2458

Wed Sep 17 2003

Review: Cognitive Science: Levinson (2003)

Editor for this issue: Naomi Ogasawara <>

What follows is a review or discussion note contributed to our Book Discussion Forum. We expect discussions to be informal and interactive; and the author of the book discussed is cordially invited to join in.

If you are interested in leading a book discussion, look for books announced on LINGUIST as "available for review." Then contact Simin Karimi at


  • Thora Tenbrink, Space in Language and Cognition

    Message 1: Space in Language and Cognition

    Date: Wed, 17 Sep 2003 00:43:49 +0000
    From: Thora Tenbrink <>
    Subject: Space in Language and Cognition

    Levinson, Stephen C. (2003) Space in Language and Cognition: Explorations in Cognitive Diversity, Cambridge University Press, Language, Culture and Cognition 5.

    Announced at

    Thora Tenbrink, University of Bremen, Germany


    ''Space in Language and Cognition'' was written by the Director of the Language andCognition department of the Max Planck Institute for Psycholinguistics (MPIP) in Nijmegen. The book draws on research carried out by MPIP members during the past decades, summing up insights on the topic of spatial thinking gained in field work in many different cultures. Specifically, Levinson is concerned with spatial frames of reference. Based on the finding that some cultures rely on absolute frames of reference both in language and cognition - in contrast to cultures like ours that mainly employ intrinsic and relative frames of reference - he addresses the controversial issue of the relation between language and thought.

    Chapter 1, The Intellectual Background: Two Millenia of Western Ideas About Spatial Thinking -- In this introductory chapter, Levinson motivates the research on spatial reference frames in relation to the fact that many basic scientific notions with regard to space are deeply rooted in Western conceptions about the world. Thinking about locations in terms of 'place' rather than 'space', for example, has an equally long tradition as a preference for 'relative' and 'egocentric' rather than 'absolute' dimensions, and it is commonly assumed that language universally reflects such basic conceptions. Many accounts of spatial cognition do not seriously allow for any far-reaching variability, being based on the facts about language and thought that are observable in the researchers' cultural home environment.

    In contrast, Levinson aims at pointing out systematic differences in spatial thinking that were previously excluded or neglected by experts because of a pervasive bias towards Western cultural traditions. Thus, the fact that Western languages make heavy use of the human body for spatial notions (as pointed out, e.g., by Clark 1973) does not prove that this is universally true. Spatial relationships expressed in terms of one (e.g., relative) reference frame cannot be translated without further information into a different reference frame (e.g., absolute). Since the languages of the world differ profoundly in the range of reference frames they provide, it is clearly premature to draw conclusions from the facts of only one kind of language about any innate or universal concepts they are assumed to reflect.Levinson maintains that no direct one-to-one mapping between non-linguistic concepts and the semantics of linguistic expressions should be assumed, but points out that, with regard to reference frames, fundamental diversity is not only found on the linguistic level but also in non- linguistic codings of spatial scenes.

    Thus, although language cannot be regarded as a direct window to cognition, distinctions expressed in language clearly need to be supported by cognition. His conclusion (from the facts presented in the later chapters of the book) is that ''human spatial thinking is quite heavily influenced by culture, and more specifically by language; when languages differ in crucial respects, so does the corresponding conceptualization of spatial relations'' (p 18). That this is a non- trivial finding is obvious through the fact that much scientific research is concerned not only with spatial concepts and language, but also with other domains, such as time, that are generally assumed to be based on, or at least closely related to, the domain of space.

    Chapter 2, Frames of Reference -- Much of this chapter draws on a previous, well-received publication of the author, in which he develops his conception of spatial reference frames (Levinson 1996). First, he points out that - in contrast to some assumptions in the literature - it is not the objects themselves, or their change of location, that constitute the differences between different ways of perceiving spatial relations, but rather the underlying abstract ''coordinate systems''. Then, he summarizes diverse accounts of spatial reference frames in the literature, points to numerous contradictions and confusions of terminology, and goes on to propose his own classification. He distinguishes three basic kinds of frames of reference: intrinsic, relative, and absolute. This account cross-cuts many other distinctions previously drawn and re-defines several conceptions that could not be properly captured using oppositions like 'egocentric' vs. 'allocentric', or 'deictic' vs. 'intrinsic'. Levinson's intrinsic frame of reference is based on a binary relation between referent and relatum (identical to origin), while the relative frame of reference uses a ternary one between referent, origin, and relatum. In both frames of reference, the origin can be the speaker, the addressee, or a third entity. The widespread term 'deictic' is not - in Levinson's view - a sensible characterisation for spatial frames of reference since it confuses binary and ternary relations, which are logically different. Absolute frames of reference rely on arbitrary fixed bearings, 'cardinal directions', which can be related to compass bearings (although they usually do not correspond directly to them). Using such a system requires maintaining one's orientation with respect to the fixed bearings at all times. As Levinson shows in detail in later chapters, people using languages that rely on absolute frames of reference are in fact permanently aware of their 'absolute' orientation, although it is not clear how they do it, i.e. how it is possible that people use an 'internal compass' that people in Western cultures simply do not have available. Obviously, such people need to maintain a constant background calculation of cardinal directions, no matter where they are, whether or not they are in familiar surroundings, and in which direction they are oriented.

    Chapter 3, Linguistic Diversity -- In this chapter, Levinson presents in more detail how the semantic parameters outlined in the previous chapter are selected and arranged in different languages. An overview shows the ways in which a ''Where''-question (the availability of which apparently is a language universal) can be answered by either employing a frame of reference or by using placenames, deictic expressions and gestures, or by relying on contiguity or topology. In this classification, 'deictic' expressions are not used in frames of reference but rather as a means of providing landmarks, such as 'here' and 'there', yielding radial specifications without an underlying coordinate system. The three frames of reference can also be applied to the vertical dimension, which does not play a major role in this book, but is briefly outlined in the present chapter. Furthermore, the frames of reference are differentiated into further subgroups, including some observations with regard to the area of motion, and some information about distributional patterns across languages is presented.

    Chapter 4, Absolute Minds: Glimpses into Two Cultures -- This chapter provides some deeper insight into two cultures (Hopevale and Tenejapa communities) where absolute frames of reference play a major role. Levinson presents fascinating anecdotes as well as detailed statistical results from the investigation of both linguistic and non-linguistic features in field work carried through by himself and other MPIP researchers. The methods used are described in some detail, and it is pointed out in which ways the investigations can be said to be significant or still require continuation or improvement. The analysis shows how the two communities differ in some crucial respects and still share fundamental cognitive properties based on the fact that they both rely on absolute frames of reference in language and cognition.

    Chapter 5, Diversity in Mind: Methods and Results From a Cross- linguistic Sample -- In this chapter, cross-cultural investigations are examined in detail testing - and confirming - the hypothesis that there is a correlation between linguistic and non-linguistic codings of comparable spatial scenes. After establishing the details of this correlation, Levinson rules out further possible determinants other than language itself that could bring about this parallelism. His conclusion is that ''populations converge on a particular non-verbal coding strategy largely because they have learnt to do so by communicating with each other'' (p 213), i.e., language is the driving force that leads to the different cultural strategies with regard to the choice of frame of reference.

    Chapter 6, Beyond Language: Frames of Reference in Wayfinding and Pointing -- While previous chapters concentrated on language and non- linguistic coding abilities such as those needed in memory and reasoning, this chapter addresses two further areas reflecting aspects of spatial cognition, namely, wayfinding and gestures during speaking. It turns out that people living in cultures relying on absolute frames of reference are specifically - and astonishingly - good at finding their way home even in unfamiliar surroundings, as well as at estimating directions even to places where they have never been before. This implies that these people integrate information about the location of places into a mental map (without ever using a real one), including angles and distances in relation to other places. This kind of mental map is fundamentally different to the kind of 'strip-map' (Tolman 1948) readily available to people in cultures utilising primarily relative frames of reference. This amazing ability of 'absolute speakers' seems to be imparted through language: when talking, these people use gestures indicating absolute directions even in telling narratives, thereby reflecting correctly the orientations and spatial arrangements referred to in their story. Such gestures can point in all directions, even 'through' (i.e., behind) the body.

    Chapter 7, Language and Thought -- The last chapter can be characterised as an empirically motivated philosophical essay about the nature of the relationship between language and thought, and about the implications with regard to theories of conceptual structure as well as the Whorfian hypothesis of linguistic relativity. Levinson contends that, in contrast to claims that language merely reflects underlying concepts, language facilitates cognitive development, making available the concepts to be developed, and permanently (re)structuring cognition. A brief look into language acquisition data supports this hypothesis via the fact that no frame of reference seems to be innate; none is acquired before the age of four, and which frame of reference is acquired first depends on the predominant one in the respective culture. Thus, fixed arbitrary bearings equivalent to 'north' are acquired by children in 'absolute' cultures between the ages of four and six - the same age at which Western children acquire the supposedly innate intrinsic frame of reference that our languages rely so much upon. In a comprehensive overview (p 314), Levinson proposes a list of universals of frames of reference in language and associated cognition, such as: ''All combinations of frames of reference are possible in a language, except that a relative frame of reference implies an intrinsic one''. This list exemplifies the fundamental contrast between Levinson's findings and such widespread (Western traditional) ideas as those recapitulated in Chapter 1.


    This book is a highly valuable accomplishment in several respects: It summarizes the findings of many years of research in various different cultures carried out by a number of MPIP field workers in an interesting and highly readable way. It disproves several popular hypotheses and traditional assumptions with regard to spatial language and concepts. It presents a comprehensive overview over the range of variability in spatial reference and non-linguistic spatial concepts, and it goes beyond the restricted domain of spatial cognition to contribute to scientific debates momentous to many different areas of research in various disciplines. In addition to the book's own virtues it should be pointed out that field research on the lesser known languages and cultures of the world is a tremendous achievement in itself. To work out such detailed and well-informed results as presented in this book (as well as, of course, in the many other publications of the MPIP and other institutions) requires years of commitment to the hardships of adapting to a fundamentally different culture. Clearly, experiments carried through only with the help of an uninformed interpreter (which is often done to avoid the many problems associated with extensive field work) could never provide nearly as much insight.

    In the following, I offer some ideas and remarks that hopefully give rise to further engagement with the topic.

    In his overview of spatial reference in Chapter 3, Levinson concentrates on expressions that could answer the question as to ''where'' an object (or a place) is located. His overview contains also descriptions of other kinds of spatial reference, but - similar to most other research in this field - it neglects one broad issue, namely, how objects are differentiated from other (similar or different) objects using qualitative spatial expressions (cf. Moratz et al. 2001), answering a question such as ''which object (do you mean)?''. As it turns out in our own experience in human-robot interaction (where this kind of 'group-based' reference is a useful way of mediating between the interlocutors' perceptions), this issue is not trivial in that there is no one-to-one correspondence between the frames of reference as worked out by Levinson and group-based strategies of referring.

    Also research findings addressing further issues are largely left out of Levinson's overview, such as those concerning application (or acceptability) areas in comprehension and production (e.g., Zimmer et al. 1998), the question how linguistic modifications and combinations of referring expressions are used for non-prototypical reference areas (e.g., Vorwerg 2001), or alternative methods of localising such as the employment of quantitative rather than qualitative expressions, specifying angles and distances, or relying on an object's features, which are all pervasive in spatial communication. Levinson also does not pursue the fact that an absolute frame of reference - using compass directions - does indeed serve several communicative functions (not only in external maps!) in Western societies. That communicative functions can play a major role in the choice of frames of reference is shown by Senft (2001) for absolute and relative frames of reference in Kilivila. This issue could clearly have received more attention in this book (Levinson mentions the matter briefly when describing situation- specific vs. unspecific experimentation). Thus, Levinson's intuitively appealing characterisation of frames of reference should be viewed as a (useful) simplification of a complex area of research. Of course, considering the book's aims it stands to reason that one must abstract from some details in order to be able to grasp the overall pattern that Levinson develops convincingly.

    In spite of the detailed theoretical definitions of the three frames of reference it is not always transparent how, in a given situation, the relevant reference frame should be determined. Take Levinson's example (p 270) in which ''the knife is to the right of the fork'': the same configuration is achieved by employing an intrinsic reference frame using a person seated at the table as origin, so that the fork is on the left of the person, while the knife is on the right. It stands to reason that people's good memory of scenarios like this rely more on the everyday functional application (using an intrinsic frame of reference) than on a context-free conceptualization of a knife being located to the right of a fork (using a relative frame of reference). Similarly, the comparison of gestures, wayfinding abilities, and route descriptions of 'absolute' and 'relative' speakers suffers from the fact that route descriptions mostly rely on an intrinsic reference frame based on a 'generic wanderer' (Herrmann & Grabowski 1994) rather than a relative one, and gestures are equally likely to rely on the speaker's intrinsic features rather than on a third entity (in Ch. 6 Levinson offers a distinction between intrinsic and relative reference frames reflected in gestures which, to my mind, needs further motivation).

    Concerning the Whorfian hypothesis, Levinson seems to pursue the idea that communication shapes thought to a certain degree; however, mostly his emphasis is on the role of language rather than language use. That this is a fundamental difference becomes clear considering the century- old distinctions between langue and parole, or competence vs. performance, etc. A terminological confusion (if it is only such) could therefore lead to some major misconceptions.

    My observations are in no way intended to diminish the author's accomplishment in presenting this book. It is strongly recommended to all researchers and students of spatial cognition and beyond, presenting both new ideas and a framework that is relevant to a broad range of sub-areas in the field.


    Clark, H. H. 1973. Space, time, semantics, and the child. In T.E. Moore (ed.), Cognitive development and the acquisition of language, pp. 28- 64. New York: Academic Press.

    Herrmann, T. & J. Grabowski. 1994. Sprechen. Psychologie der Sprachproduktion. Heidelberg: Spektrum.

    Levinson, S. 1996. Frames of reference and Molyneux's question: Cross- linguistic evidence. In P. Bloom, M. Peterson, L. Nadel and M. Garrett (eds.), Language and space, pp. 109-69. Cambridge, MA: MIT Press.

    Moratz, R., K. Fischer, & T. Tenbrink. 2001. Cognitive Modelling of Spatial Reference for Human-Robot Interaction. International Journal On Artificial Intelligence Tools, 10:4, World Scientific Publishing, Singapur.

    Senft, G. 2001. Frames of spatial reference in Kilivila. Studies in Language 25:3, 521-555.

    Tolman, E.C. 1948. Cognitive maps in rats and men. The Psychological Review 55(4):109-45.

    Vorwerg, C. 2001. Raumrelationen in Wahrnehmung und Sprache: Kategorisierungsprozesse bei der Benennung visueller Richtungsrelationen. Wiesbaden: DUV.

    Zimmer, H.D., H.R. Speiser, J. Baus, A. Blocher, & E. Stopp. 1998. The Use of Locative Expressions in Dependence of the Spatial Relation between Target and Reference Object in Two-Dimensional Layouts. In C. Freksa, C. Habel & K.F. Wender (eds.), Spatial Cognition. An Interdisciplinary Approach to Representing and Processing Spatial Knowledge (pp. 223-240). Berlin: Springer-Verlag.


    Thora Tenbrink is a research assistant in the newly established DFG Collaborative Research Center SFB/TR8 "Spatial Cognition: Reasoning, Action, Interaction" (Bremen & Freiburg, Germany). Her research interests focus on the fields of discourse analysis and text linguistics; previous work has dealt with discourse relations and information structure, presuppositions and non-temporal implications of temporal connectives, especially 'before' and 'after'. Her current focus is on empirical research on spatial reference systems in human- robot interaction. Her dissertation project deals with discoursal applicability conditions and features of spatial (and temporal where applicable) expressions in human-robot interaction.