Ancient Northwest Semitic languages (NWS) have hardly been included in the studies of the linguistic typology of spatial relations so far. The other way round, areal linguistics of the languages of this dialectal continuum have not yet reached the level of semantics. This Ph.D. project contributes to close these gaps, focussing on the static subsystems of NWS contrastively: with each other, with German (translational target language) and with the typological research.


Spatial relations attested in Ugaritic nonverbal clauses with their German equivalents: convergence and differences in the semantic (and hence cognitive) segmentation of space | Illustration: Jörg Hartlieb

(Click to enlarge) Spatial relations attested in Ugaritic nonverbal clauses with their German equivalents: convergence and differences in the semantic (and hence cognitive) segmentation of space | Illustration: Jörg Hartlieb

Space enters focus in life, hence cooperation and hence communication, when it becomes relevant in a situation. Since the means for efficient human communication are human languages, they provide and form semantic-pragmatic and morpho-syntactic subsystems for referencing space, with the former being reconstructible via the latter. When the location of a participant (or situation) is in focus – and not its inherent or internal spatiality – linguistic localization relates it (the locandum or figure) to a (different) concrete spatial participant (the locatum or ground) that is on-hand and invariable in the communication situation. Such spatial relations can be macro-categorized, based on the oppositions (a) of motion and stasis, within the latter (b) of salience and ambiguousness of the relata involved, and within the latter (c) of intrinsic, relative and absolute frames of reference to relate the locandum to the locatum linguistically (and cognitively). Further sub-categorization is highly variable even with closely related languages (for example, while in English the apple is “on” the spike and the ring is “on” the finger just as a jar is “on” the table, Dutch distinguishes these relations: aan, om, op).

Theoretical framework

Since language is functional in essence, a functional-typological theory of language based on Thomas Givón (2001) is chosen as unifying framework to reconstruct and integrate previous research and new results. The functional relationship of space, cognition and language is outlined, resulting in a functional definition of ‘spatial relation’ and a mapping of its relata to semantic roles, their semantic roles to grammatical roles and hence spatial relations to their typical sentence patterns. Grammatical roles or philological categories often imposed to NWS from the language of (the) research(er) are redefined by functionally establishing the respective morpho-syntactic criteria.

Since there is no objective and sufficient access to human cognition (in particular without living native speakers), meanings of linguistic constituents are defined by reconstructing their conditions of usage from their linguistic and non-linguistic context (operational semantics, following Ernst Leisi 1975 & 1985).

To simplify integration into the previous typological research, tools of Levinson and his research group are adapted, in particular the Topological Relations Pictures Series. Hereby the basic grid of sub-categories reflected in them is adopted. Since, however, those tools are non-linguistic stimuli in nature, translations of them into semantic features are proposed to facilitate explicit semantic definitions of sub-categories – a desideratum in the previous research, which partly (Levinson) or fundamentally (Daniel A. Werning (2013) on Egyptian, Ulrike Steinert (2013) on East Semitic Akkadian) rely on a purely informal notion of resemblance to the stimuli.


To simplify interdisciplinary accessibility the Leipzig Glossing Rules are adapted for NWS with different levels of reconstruction (transliteration vs. transcription) and morphemic abstraction. Intrinsic to Levinson’s tools is the visualization of results in semantic maps.


Spatial relations can be identified in ancient corpora by identifying the relata (or ‘similar’ ones) pictured on the stimuli, by searching for verbal meanings with prototypical sentence patterns, by analyzing whole (sub)corpora for localizations and by searching for supposed morpho-syntactic means, e.g. prepositions.

The first strategy (successful with Werning’s and Steinert’s corpora) proved to be less appropriate for NWS, since even for basic relations the respective relata are lacking. The last strategy is problematic, since it predetermines the range of results. A general problem is the lacking of a scholarly consensus as to semantics, grammatical analysis and textual analysis as well as of up-to-date (and electronic) tools to access it. Hence the third proved to be the most appropriate strategy, despite its effort.

Current state

The language under analysis still is Ugaritic. The figure shows as example the visualization of topological relations according to the sub-corpus of Ugaritic non-verbal clauses.

This Ph.D. thesis is being written within the program “Ancient Languages and Texts” (ALT) of the Berlin Graduate School of Ancient Studies (BerGSAS).