• Dialetti Beduini
• Dialettologia Araba
• Frames of Reference
• Space Cognition in Arabic Dialects

Object-Based Selection of Spatial Frames of Reference in aṣ-Ṣāniˁ Arabic

Introduction
Levinson (2003) defines spatial Frames of Reference as semantic and cognitive strategies used to project coordinate systems onto spatial arrays; by such means, we can conceptualize and linguistically describe projective (or angular) spatial arrays (Tables 2.1 and 2.2). They are basically three across languages: Intrinsic, Absolute and Relative (Tables 2.3, 2.4, 2.5, 2.6). Starting with the preliminary observation that all three Levinsonian Frames of Reference occur in the Bedouin dialect of aṣ-Ṣāniˁ elderly speakers in the Negev, the present research attempts to account for the strategies underlying the selection and distribution of the Relative Frame of Reference. I focus in particular on the ALIGNED FIELD, a sub-category of the Relative Frame of Reference, first detected in the Hausa language (Hill 1982). I restrict my field of investigation to the linguistic level of analysis, without addressing the question of the relationship between language and cognition. In fact, within the domain of space, Frames of Reference (as well as prepositions and verbal properties) have been considered very enigmatic and disputed between supporters of Relativism (Humboldt 1936; Boas 1911; Whorf 1956; Talmy 1978, 1988; Peterson et al. 1996; Slobin 1996; Levinson 2003; Evans & Levinson 2009) and those of Universalism (Chomsky 1965; Comrie 1989; Whaley 1997; Croft 2003).
Hypothesis
With respect to the distribution of Frames of Reference within a language, Tversky and Levinson proposed different interpretations of this phenomenon. The first author, operating in the frame of cognitive studies, observes:
‘(…) it is a common enough finding that in spatial description in some languages alternative frames of reference will be available, and which one is actually employed will depend on properties of the task – e.g. the scale of the things to be described, what use the information is to be put to, and so on’ (Tversky 1996).
Levinson (2003: 179-181) attempts to delve into typological classification of languages and to test the ‘language-to-cognition correlation’. He writes:
‘(…) we need to distinguish between (a) a cross-situation typing, (…) [based on] a general preference for one frame of reference over another across situations, and (b) a typing that is keyed to a specific kind of spatial array or situation (…) most languages provide special expressions for more than one frame of reference, and there are conventions for the kinds of circumstances each frame of reference is used in. So we need to relativize the statement to situations of use. (…) The idea behind the hypothesis is that community-wide conventions about what linguistic expressions mean and how they are to be used will tend to induce a way of thinking in which the immediate, unreflective memory coding matches the kind of coding required to describe an arbitrary spatial array’.
Therefore, my initial question was: What are the ‘speech circumstances’ or rather ‘community-wide conventions’ determining the selection of the Relative Frame of Reference on the purely linguistic level in the Arabic dialect of aṣ-Ṣāniˁ elderly speakers? The answers of my informants led me to concentrate on the second aspect, ‘community-wide conventions’ for the selection of spatial Frames of Reference, even though ‘speech circumstances’ could be a promising field of enquiry on the selection of spatial Frames of Reference as well, since cross-generational, sociolinguistic and stylistic variations in Negev Arabic dialects have been detected and described in the methodological framework of sociolinguistics and discourse analysis (Henkin 2010).
Methodology
Informants
Because of its interesting and rich set of semantic strategies describing spatial relations, because of the change from nomadic to sedentary life style and for its genetic, cultural and territorial conservatism, I chose to conduct my fieldwork on the aṣ-Ṣāniˁ ˁišīrih, small tribe or rather large family of Bedouin inhabiting the northern borders of the Negev desert. I divided the informants in three age-groups, comprising male and female elements, representing three fundamental steps in the history of the aṣ-Ṣāniˁ community, within the larger frame of the history of the Middle East and of the State of Israel, i.e. three fundamental steps along the path of cultural change. The investigation of the language of space among the elderly aṣ-Ṣāniˁ in comparison to the language of the younger generations yields the unprecedented opportunity to observe and to typologically describe two distinctively recognizable stages in the history of the language of a community which has maintained itself as a homogeneous social entity throughout the challenging events of the last century, until the abruptly accelerating acculturation process of the last two decades.
Stimuli
To typologically investigate a language requires a deep analysis of its semantics, rarely available from grammar books. The complexity of this enquiry mostly consists in devising special methods of data elicitation – such as, for example, communication tasks between native speakers and pertinent sets of controlled stimuli. ‘Pertinence’ is the key word of this methodological treatise, i.e. the consistency of methodology and stimuli design with both the scope of the enquiry and with the social and cultural context of the investigation. Therefore, the first chapter of the present thesis sets out to describe the development of a culture-specific methodology, implemented and adopted in my fieldwork. I report the results of the combination of cognitive semantics with the epistemological instruments of cultural anthropology, a noteworthy aspect of the present investigation. The very beginning of the enquiry was characterized by the cross-linguistic approach proposed by the team of the Max Planck Institute for Evolutionary Anthropology: Typological surveys on static spatial relations generally aim at matching linguistic and non-linguistic representations applying the same sets of stimuli across languages and cultures. Indeed, the consistency of the methodology applied across different languages is considered itself as the best guarantee for the validity of the data yielded by cross-linguistic enquiries (Levinson & Meira 2003; Levinson & Wilkins 2006). Nevertheless, the response of my informants induced me to a radical change in methodology, especially of the design of the stimuli (layouts, materials, objects): The closer the stimuli were to the traditional culture of the elderly, the more the peculiar features of the older language were revealed. Furthermore, the reactions and answers of the younger informants, product of a new and different material culture, showed different linguistic expressions of spatial relations. These observations served to highlight the central role played by the ontological classification of the objects in the linguistic expression of spatial relations. The change in methodology was implemented gradually after subsequent sessions of my fieldwork.
STAGE I: I started showing on the screen of my computer or as photos on paper BIDIMENSIONAL INDIRECT STIMULI (representations of real objects in forms of pictures, as photos on paper), subdivided into three sub-categories:

1. culturally non-related stimuli pictures;
2. culturally related stimuli pictures (scenes taken from Bedouin life);
3. artificial stimuli pictures (experimental arrays in non-arbitrary positions, i.e. selected to detect specific distinctive features) (Table 1.3).

The images, shown on the computer screen or on paper, were often not recognized as the bi-dimensional representations of tridimensional entities and, therefore, were not processed according to the geometric and ontological semantic rules commonly used in the treatment of concrete spatial arrays. They were simply treated as bi-dimensional objects: Measures, proportions, distances and relative positions of objects in pictures were not always recognized as clues of perspective rules, to be applied by default to properly read images on bi-dimensional layouts (Table 1.4). In the case of pictures shown on the computer screen, the vertical dimension according to which the screen is oriented was chosen to describe the represented array (Table 1.5) or, in the case of unshaped objects, the informants resorted to the Absolute Frame of Reference (Table 1.9). Unknown geographical elements were not recognized and not described by the elders (Table 1.7), while known mountains and rivers were selected as Ground objects with their real orientation (Table 1.8) and schematic representations of mountains and rivers were paradigmatically associated with the concrete local experience of the same real natural elements and were oriented accordingly (Tables 1.10 and 1.11).
STAGE II: Since BIDIMENSIONAL INDIRECT STIMULI seemed to be somehow misleading, I resorted to the use of tridimensional real objects or toy-objects. I noted that toy objects, tridimensional representations of real entities were recognized without difficulty as representing the original entity (man, horse, tree etc.) on scale and the reduction did not adulterate the ontological properties of the real entity itself in spatial discourse (for discussion, Danziger & Gaskins 1993). So, I divided the new TRIDIMENSIONAL DIRECT STIMULI into three categories:

1. Culturally-Related Objects (tent, knife, horse, donkey, camel, sheep, goat, dog, wild carnivore, tree, stone, fence, coffee pot, cup, fireplace);
2. Formerly-Acquired Objects (agricultural products);
3. Recently-Acquired Objects (house, car, phone).

So, according to the aforementioned principles of stimuli-categorization, as recommended by Danziger & Gaskins (1993), I created my own superset of stimuli and games calculated to highlight distinctions relevant in traditional aṣ-Ṣāniˁ Arabic, in the most efficient and most exhaustive way possible. I started the new cycle of sessions by testing the most cross-linguistically diffused distinctions, using objects selected as Ground, which should present themselves with and without the following distinctive features:

1. Animacy
2. Personhood
3. Capability of motion
4. Facedness (without implied motion)
5. Symmetry
I gradually individuated more culturally-peculiar semantic distinctions.

Procedure

This work presents the answers given by the aṣ-Ṣāniˁ informants to where-questions about the position of a certain object (Figure) with respect to a second one (Ground) by means of different kinds of controlled stimuli. The aim of the investigation is to outline the semantic treatment of the projective spatial relations on the horizontal plane, i.e. of those spatial relations between entities whose description requires – in all languages worldwide – the use of some kind of coordinate system (Frames of Reference).
In order to test this kind of spatial relationship, the routine where-question accompanying every spatial array or image was: ‘where is X-object with respect to Y-object?’, wīn X min Y?. To impose a preliminary selection of Figure and Ground using the structure of the question ‘where is X with respect to Y?’ is a quite artificial but necessary practice to accomplish the aims of the present research, where a well-defined set of spatial relations had to be investigated: In natural speech situations, describing a spatial array by giving topological or projective information is most of the time a personal choice of the speaker, due to a number of different particular situation-based priming causes.
First of all, before every session of fieldwork, I ascertained that the informants recognized all the objects which were going to be used, so we agreed on the words. I divided every session of my fieldwork into thematic sequences. Within every thematic sequence I changed the objects of every array, in order to avoid responses containing sequential topological information, like: ‘X has come closer to Y’, ‘X has gone away from Y’, ‘X has moved to the other side of Y’ and so on, and to shield the informants from the tendency to create a story out of consecutive arrays, i.e. to take a former scene as reference point for the description of the subsequent array. I also changed my own position often during the session, repeating the same questions on given arrays many times, while assuming different positions. Indeed, I noticed that given two objects, X Figure and Y Ground, independently of the framing system used by the speaker, the presence of additional objects or people around the array, mostly if these are intrinsically oriented, can affect the outcome of the experiment (Table 1.13).
The second part of the methodological explanations is presented in the preliminary section of the second chapter, in the form of a general overview of the basic means used to explore the domain of space in the typological analysis: concepts of Ground, Figure, Projective Relation, Region and Frames of Reference are introduced before the typological classification of the Frames of Reference and their relevance to the debate on the existence of universal categories in language and cognition.
In order to test the presence of the Relative Frame of Reference in the language of aṣ-Ṣāniˁ elderly speakers, the nucleus of the present research, and in particular of its effects on the projection of the Front / Back Axis of the Ground objects, I selected stimuli with different kinds of Ground objects: a stone, a tree, a flock, a donkey, a horse, a man, a sheep and a goat. All these objects are part of the speakers’ traditional cultural environment. Some of them could easily attract the projection of external coordinate systems, being geometrically ‘symmetric’, i.e. showing no inherent difference along the Front / Back Axis nor on the Right / Left Axis, like stones and trees. Ground objects like stones and trees are largely considered as inanimate symmetric objects and largely used in the literature for detecting the Relative Frame of Reference (see: ‘Man and Tree Stimuli’ in Levinson 1992, kit 1.: 7-14 and Hill 1982: 22).

Results
As a preliminary result, Table 2.7 presents the contrastive distribution of Relative and Intrinsic Frames of Reference based on the type of Ground object: donkey / horse / man as Grounds prime the Intrinsic Frame of Reference (images 4, 5 and 6), while no salient asymmetries are inherently recognized by the speakers to stone / tree / flock on the Front / Back Axis. Such asymmetries are projected from the Observer onto stone / tree / flock as Grounds, set in the middle of the visual field of the Observer, according to the rules of the Relative Frame of Reference applied by Translation (Table 2.6), as happens in the Hausa ALIGNED FIELD (Hill 1982).
Furthermore, with respect to the Hausa system, the aṣ-Ṣāniˁ system presents a noteworthy asymmetry in the use of prepositions. While wara represents the Back Region (behind) both in the Intrinsic and in the Relative Frame of Reference, the use of giddām (in front of) is restricted to the Ground objects priming the Intrinsic Frame of Reference (Table 2.8). The Front Region in the ALIGNED FIELD is variously described. In other words, given a symmetric Ground and the geometric circumstances of an ALIGNED FIELD, the projection of the Front Region and of the Back Region onto the Ground object do not occur by default, as seems to be the case in the Hausa system. The aṣ-Ṣāniˁ speakers look at the properties of the Ground objects rather than following the geometric layout of the visual field. Projecting a Front Region onto a stone / tree / flock or onto a single sheep / goat is thus perceived by the aṣ-Ṣāniˁ speakers as stranger or erroneous than projecting a Back Region onto them, since the Back Region is prototypically less animated than the Front Region in the asymmetric and mobile bodies, from which such a semantic distinction originally emerged. Furthermore, the semantic origin of wara (adverbial root) is literally less ‘embodied’ than the etymological origin of giddām, related to the body part noun gadam (*qadam) ‘foot’. In an ‘object-sensitive’ spatial system, this fact can favor the extension of wara (and other prepositions not originally bound to body-parts) to a larger set of ontologically differentiated entities (with respect to properties like [ANIMACY], [DIRECTIONALITY] and [MOBILITY]) and to a larger set of geometric arrays than the set of situations where it is possible to properly use giddām. Therefore it seems that the aṣ-Ṣāniˁ language preserves the remnants of a spatial system deeply based on the features of the Ground objects protagonists of the arrays. We see indeed, that, even in the application of the ALIGNED FIELD, which is a sub-category of the Relative system, the characteristics of the objects and the values they possess with respect to the criteria of [ANIMACY], [MOBILITY], [DIRECTIONALITY] have a priming effect in the selection of strategies for the description of their spatial relations.
In a cross-generational perspective, the general tendency among the younger generation (adults, not old) is the use of the preposition giddām to express the Front Region of a stone / tree / flock in the ALIGNED FIELD. To understand the new system, I can hypothesize a shift of attention from the inherent properties of the objects to the geometric oppositions of the arrays. In other words, younger people use giddām because of the geometric opposition giddām-wara (‘in front of’-‘behind’).
Interestingly, the importance of the properties of objects does not entail an increased use of the Intrinsic Frame of Reference in the language of the elderly: on the contrary, the Intrinsic Frame of Reference is widely observable in the language of the young. This may be due to the prevalence of functional-intrinsic criteria for classifying many objects of modern life; it may also be related to a more restricted and localized living space in the new sedentary life style, and multiple intrinsic restrictions on moving about in this space, such as roads and traffic signs.
The primacy of the culturally attributed properties of the objects in the way the elderly speakers describe the spatial relations represents the strategy of selection of the more appropriate Frame of Reference among the large asset of possibilities that they exploit. In the case of a sheep / goat Ground object, the ALIGNED FIELD is realized using two prepositional phrases: minnih w jāy ‘from it and towards me’ / minnih w ġād ‘from it and away’, regardless of the direction the sheep / goat is facing. The only exception is coincidence of the rear part of the sheep / goat with the Back Region of the ALIGNED FIELD (Table 2.11 image 1). I consider these prepositional phrases as distinctive markers of the ontological properties of the sheep / goat as Ground objects, distinct from other symmetric objects like the stone / tree / flock.
But these two prepositional phrases are also used in other situations where the Relative Frame of Reference is preferred, such as the case of the Figure object partly or totally hidden by the Ground object: minnih w ġād is used is used to indicate the position of a Figure totally hidden by a Ground, instead of the use of baya ‘behind’ shown in the Hausa language (Table 2.9). The doublet minnih w jāy / minnih w ġād is used for the sheep / goat also in all those cases where the anatomical part of the animal can conflict with the Region of the ALIGNED FIELD, like in the case where the animal is facing toward the Observer, i.e. the face of the animal corresponds to the Back Region of the ALIGNED FIELD. In such cases wara is felt to be improper (because of the ‘light anatomical effect’ of the snout of the animal), as is giddām (because of the restrictions of the ALIGNED FIELD) (Table 2.13). In other words, when the inherent properties of the entity in the domain of space are weak and produce ‘light’ effects, the geometric rules of the visual field of the Observer tend to overrule the ontological properties.
As shown in Table 2.16, the effects of the ALIGNED FIELD are evident also when the Observer sees the sheep / goat from a side, i.e. the intrinsic Front/Back Axis of the sheep / goat should be perpendicular to the direction of the ALIGNED FIELD and this fact should generate a conflict of attribution of Regions. In those cases, as an Italian speaker, I would expect something like ‘l’albero è accanto alla pecora’, ‘the tree is beside the sheep’, according to the Intrinsic Frame of Reference, recognizing in the sheep / goat its lateral side. But this is not the case: similarly to what happens in the case of the telephone in Hausa, aṣ-Ṣāniˁ speakers resort to the use of the Relative Frame of Reference, using the doublet minnih w jāy / minnih w ġād that we have observed in Table 2.10 image 2 (the same case as in Table 2.11 image 2) and Table 2.12 images 1 and 2. It means that in the domain of space, the sheep / goat as Ground objects, like the telephone in Hausa, show no difference between their Front / Back Axis and Right / Left Axis, similarly to what happens with the stone / tree / flock as Ground object. Similarly to the case of the Hausa telephone, it does not matter where the sheep /goat is facing, since it is treated as a symmetric Ground in all its facets.
The collapse of the ALIGNED FIELD among the young generations is evident from the fact that the opposition wara / some geometric and metric property (as we already know) is substituted by gabl (before) / baˁd (after) to lexicalize the opposition Back Region / Front Region of the ALIGNED FIELD. Significantly the use of these prepositions is otherwise restricted in aṣ-Ṣāniˁ exclusively to the temporal domain. The use of gabl (before) / baˁd (after) in the ALIGNED FIELD implies that when the Figure object is between the Observer and the Ground, the young Observer would say that ‘F is gabl G’ and when the Figure is on the other side of the G, the young Observer would say ‘F is baˁd G’. Nevertheless, in practice, this system is not always realized so perfectly, generating many afterthoughts in the informants: old mental and semantic structures make their effects felt for a long time along the generations, while the system is changing. The afterthoughts of the informants are due to the conflict in the attribution of the meanings of Anteriority and Posteriority when describing the Back Region of symmetric ground-objects. When the Figure stands between the Observer and the Ground the young tend to see Anteriority and use gabl ‘before’ whereas the elderly see Posteriority and use wara ‘behind’.
Conclusions
From Table 2.9 and the observations following it, we see that the ALIGNED FIELD in Hausa and aṣ-Ṣāniˁ Arabic shows noteworthy discrepancies. In Hausa the prepositions indicating Front and Back Regions are always gaba / baya respectively, independently of the object; but in aṣ-Ṣāniˁ Arabic different Ground objects, such as the stone / tree / flock and the sheep / goat, on the basis of their ontological properties, produce grammatical distinctions even in the application of the same strategy of Translation (ALIGNED FIELD).
This fact suggests first a discrepancy in the aṣ-Ṣāniˁ system between the domain of linguistic semantics (linguistic description) and that of cognitive semantics (cognitive attribution of the Frames of Reference). Indeed, the correspondence of one Frame of Reference to many linguistic strategies ends up destabilizing the belief in the correspondence of linguistic and non-linguistic (or cognitive) knowledge, accepted by Levinson still in 2003, like a residual component of the Universalist position.
Secondly, it indicates the necessity of different methodological approaches, oriented toward the analysis of the cultural ontologies of entities in the domain of space. I use the label ‘ontology’ and not ‘hierarchy’ because in fact the partition of the realia on the basis of certain properties does not entail in all languages and cultures a hierarchical classification based on a vertical scale of implications, similar to the taxonomical system according to which western contemporary cultures represent the natural world, after the evolutionary theory of C. Darwin. So, even though the sheep, the goat and the horse should belong to the same animal category on the basis of certain shared features and on the basis of our scientific precognitions, in aṣ-Ṣāniˁ Arabic they present themselves in the linguistic description as very different entities. Sheep /goat are in an intermediate position between stone / tree / flock and donkey / horse / man. One possible reason for this is that sheep / goat can be considered lower in [MOBILITY] than carnivores, horses and donkeys – sheep rarely raise their heads when moving to the ‘next tuft of grass’; they need to be herded and do not show any large scale volitional mobility.
A further interpretation of the aṣ-Ṣāniˁ treatment of sheep / goats as Grounds, intermediate between fully symmetric and fully asymmetric objects, could be promoted by the routine position of these animals with respect to humans, since they actually proceed aligned in front of the shepherds, so the back region is salient to the Observer and thus deemed worthy of linguistic marking; in contrast, the canonical positions of horses (ridden by men) and dogs (walking beside them) tend to coincide or fuse with the perspective of the speaker (refer to discussion p. 122).
The ontological properties of the objects and their partition into spatial REGIONS affect not only the Intrinsic Frame of Reference, but the entire set of Frames. Therefore, I accept Levinson’s methodological suggestions: he recommends keeping the Ground object conceptually separate from the Frame of Reference, abandoning the traditional and reductive subdivision into ‘deictic’ and ‘intrinsic’, ‘egocentric’ and ‘allocentric’ frames (Levinson 2003: 53). Instead, he introduces the three terms Ground object, Center of the axes and Anchoring point (refer to p 101) and shows how, in these terms, the sheep / goat is distinct from the stone as a Ground object, although both prime the ALIGNED FIELD (Table 2.14).
Interestingly, the ontological properties attributed to entities in the domain of space are not universally valid in all other linguistic domains: in the domain of space the sheep / goat is less mobile and, somehow, less relevant than the donkey / horse / man, but in the domain of color terms, sheep, goats and stones enjoy the same abundance of specific terms as camels and horses (Borg 2007).
In summary, aṣ-Ṣāniˁ Arabic emerges as utilizing a strategy of selection of Frames of Reference grounded on the culture-based ontological classification of entities in space, supporting a neo-relativistic approach to space studies (Marotta 2010). Even though geometrical and ‘logical’ rules still represent the methodological foundation of many theories of space categorization (Cooper 1968; Leech 1969; Bennett 1975; Miller & Johnson Laird 1976), the most evident limit of this model is the fact that, actually, Ground objects can be differently conceptualized not only on a cross-linguistic and cross-cultural perspective but also within a single language, as this study clearly shows. Some semantic theories in the 1980s already recognized this fact, representing a turning point in the progress of the neo-relativistic perspective (Jackendoff 1983; Vandeloise 1986), to which this dissertation hopefully contributes.
References
Bennet, D.C. (1975) Spatial and Temporal Uses of English Prepositions: An Essay in Stratificational Semantics. London, Longman.
Boas, Franz (1911). The Mind of Primitive Man. The Macmillan Company.
Borg, A. (2007) “Towards a History and Typology of Color Categorization in Colloquial Arabic”. In MacLaury, R.E., Paramei, G.V., Dedrick, D. (eds.) Anthropology of Color: Interdisciplinary Multilevel Modeling. Amsterdam, J. Benjamins, pp. 264-293.
Chomsky, N. (1965) Aspects of the Theory of Syntax. Cambridge, MA, MIT Press.
Comrie, B. (1989) Language Universals and Linguistic Typology. Oxford, Blackwell.
Cooper, G.S. (1968) A Semantic Analysis of English Locative Prepositions. Technical Report 1587, Cambridge, Bolt, Boranek and Newman Report.
Croft, W. (2003) Typology and Universals. 2nd edition. Cambridge, Cambridge University Press.
Danziger, E., Gaskins, S. (1993) Exploring the Intrinsic Frame of Reference. In Levinson, S.C. (ed.) Cognition and Space Kit (version 1.0): July 1993, Nijmegen: Max Planck Institute for Psycholinguistics, pp. 53-64..
Evans, N., Levinson, S.C. (2009) “The Myth of Language Universals: Language Diversity and Its Importance for Cognitive Science”, Behavioural and Brain Science 32: 429-492.
Henkin, R. (2010) Negev Arabic: Dialectal, Sociolinguistic and Stylistic Variation. (Semitica Viva 48). Wiesbaden, Harrassowitz Verlag.
Hill, C. (1982) “Up/Down, Front/Back, Left/Right: A Contrastive Study of Hausa and English”. In J. Weissenborn, J., Klein. W. (eds.) Here and there: Cross-Linguistic Studies on Deixis and Demonstration. (Pragmatics and Beyond, Vol. III, 2-3). Amsterdam, John Benjamins, pp. 11- 42.
Jackendoff, R. (1983) Semantics and Cognition. Cambridge, MA, The MIT Press.
Leech, G.N. (1969) Towards a Semantic Description of English. London.
Levinson, S.C. (2003) Space in Language and Cognition. Explorations in Cognitive Diversity. Cambridge, Cambridge University Press.
Levinson, S.C., Wilkins, D. (eds.) (2006) Grammars of Space. Cambridge, Cambridge University Press
Levinson, S.C., Meira, S. & The Language and Cognition Group (2003) “‘Natural Concepts’ in the Spatial Topological Domain – Adpositional Meanings in Crosslinguistic Perspective: an Exercise in Cognitive Typology”, Language 79(3): 485-516.
Marotta, G. (2010) “Why Space in Language? The Reasons for a Meeting”. In Marotta, G., Lenci, A., Meini, L., Rovai F. (eds.) Space in Language. Proceedings of the International Conference, October 8-10 2009. Pisa, ETS 73, pp. 7-25.
Miller, G.A., Johnson-Laird, P.N. (1976) Language and Perception. Cambridge, MA, Harvard U.P.
Peterson, M., Nadel, L., Bloom, P., Garrett, M. (1996) “Space and Language”. In Bloom, P., Peterson, M., Nadel, L., Garrett, M. (eds.) Language and Space. Cambridge, MA, MIT Press, pp. 553-77.
Slobin, D.I. (1996) “From ‘Thought and Language’ to ‘Thinking for Speaking’”. In Gumperz, J., Levinson, S. (eds.) Rethinking Linguistic Relativity. Cambridge, Cambridge University Press, pp.70-96.
Talmy, L. (1978) “Figure and Ground in Complex Sentences”, in Greenberg, J. (eds.), Universals of Human Language, Stanford, CA: Stanford University Press, pp. 625-649.
Talmy, L. (1988) “Force Dynamics in Language and Cognition”, Cognitive Science 12: 49–100.
Tversky, B. (1996) “Language and Spatial Cognition”, International Journal of Psychology 31(3–4): 3010.
Vandeloise, C. (1991) Spatial Prepositions: A Case Study from French, (English edition of (1986) L’Espace en Français). Chicago, University of Chicago Press.
von Humboldt, W. (1836) Über die Verschiedenheit des Menschlichen Sprachbaus und ihren Einfluss auf die Geistige Entwicklung des Menschengeschlechts. English translation in this text by Heath, P. (1988) Cambridge, Cambridge University Press.
Whaley, L. (1997) Introduction to Typology. The Unity and Diversity of Language. London-New Delhi, SAGE Publications.
Whorf, B.L. (1956) Language, Thought and Reality. Cambridge, MA, MIT Press.