Figure 1 - uploaded by Michael Spranger
Content may be subject to copyright.
Spatial setup in which two robots are communicating about objects in their environment. The spatial scene with the two robots is shown in the middle. The objects that populate the environment are tracked by visual processing systems (?). These systems distill a world model of the current spatial setup experienced by each robot. The left drawing shows the world model build by the robot to the left in the spatial setup. The drawing to the right shows the world model of the robot to the right. The world consists of three types of objects: robots (black arrows), blocks (yellow circles) and the box (blue square). Each robot is robot-1 in its own world model. The interlocutor is always robot-2. There are two yellow blocks (e.g. loc-3292). Additionally, there is a box landmark box-9 drawn as blue square. The box has a front side, marked by the extra blue line in the world models.
Source publication
This chapter presents an operational grammar for German spatial lan-guage, in particular German locative phrases, as a case study for processing distributed information. It investigates the complex interplay of syntactic phe-nomena and spatial semantics, with a specific emphasis on efficient process-ing of syntactic indeterminacy and semantic ambig...
Contexts in source publication
Context 1
... key property of the FCG search for an interpretation of such an utterance is Figure 10. Final part of the parsing search tree for the utterance "der block vorne". ...
Context 2
... problem that remains is how the information, in particular the uncertainty about the frame of reference, is spread so that this construction can distribute its decision on the relative frame of reference to the place where this information is needed to compute the region, namely, the corresponding functional unit. The solution is to apply percolation noun-unit-361 block-103 Figure 11. Transient structure before the application of the relative-region--perspective-marked construction (when parsing "der Block vorne von dir aus"). ...
Similar publications
This chapter presents an operational grammar for German spatial language, in particular German locative phrases, as a case study for processing distributed information. It investigates the complex interplay of syntactic phenomena and spatial semantics, with a specific emphasis on efficient processing of syntactic indeterminacy and semantic ambiguit...
You can find the published text version here: <https://www.uni-marburg.de/de/fb09/dsa/einrichtung/personen/wissenschaftler/kasper/publikationen-und-vortraege>The present article is an attempt towards a unified picture about some central syntactic and semantic changes within complex (possessive) noun phrases in German. Firstly, the expressive adnomi...
Citations
... As an intricate linguistic phenomenon, its metalinguistic categories are often indeterminate as well (Halliday 2009;Zhang 2015), depending on which perspective is adopted to approach the meaning of language (for example, formal, semantic, pragmatic). Indeterminacy occurs when there is a fuzziness in semantic categorization and it can be caused by various factors, be they epistemic, metaphysical, pragmatic or grammatical (Kies 1990;Trueswell, Tanenhaus, et al. 1994;Santos 1998;Williamson 2003;Niederdeppe 2005;Spranger and Loetzsch 2011;Zhang 2015). 2 The present study approaches the meaning of language from a SFL perspective, more specifically, logical meaning, but the term indeterminacy needs to be understood from the tension of the unique logical system between classical Chinese and English in the process of translating. ...
... Finally, we discuss systems which are flexible enough to acquire any kind of category system without a priori assumptions about the systems themselves. This paper is part of a larger research effort that tries to understand the basic processing principles [11], acquisition and evolution of spatial language [12] with a particular focus on linguistic variation. Here we focus on lexicon acquisition. ...
This paper discusses grounded acquisition experiments of increasing complexity. Humanoid robots acquire English spatial lexicons from robot tutors. We identify how various spatial language systems, such as projective, absolute and proximal can be learned. The proposed learning mechanisms do not rely on direct meaning transfer or direct access to world models of interlocutors. Finally, we show how multiple systems can be acquired at the same time.
... Obviously, there is more to the system than we have space to explain here. In particular, German requires the handling of complex syntactic phenomena such as morphology and complex word classes, and semantic ambiguity that fall beyond this paper (see[18,16] for a detailed discussion). ...
Each natural language phrase is evidence for a particular strategy of construing reality. One domain where this has been extensively studied is spatial language, which reveals an enormous amount of variation of conceptualization strategies both within a particular language and cross-culturally. This paper proposes a computational formalism for representing conceptualization strategies and shows how the formalism can be used to study and explain the evolution and emergence of spatial conceptualization strategies and their impact on shared grounded communication systems.
... @BULLET Many linguistic domains, such as spatial language , are known for their high degree of polysemy. By distinguishing between actual and potential values, such polysemous structures can be processed smoothly (Spranger and Loetzsch, 2011). ...
Cognitive linguistics has reached a stage of maturity where many researchers are looking for an explicit formal grounding of their work. Unfortunately, most current models of deep language processing incorporate assumptions from generative grammar that are at odds with the cognitive movement in linguistics. This demonstration shows how Fluid Construction Grammar (FCG), a fully operational and bidirectional unification-based grammar formalism, caters for this increasing demand. FCG features many of the tools that were pioneered in computational linguistics in the 70s--90s, but combines them in an innovative way. This demonstration highlights the main differences between FCG and related formalisms.
... This chapter is based on a new more recent implementation by Martin Loetzsch, Simon Pauw and Michael Spranger (Spranger et al, 2010a). The current implementation has already been used in language game experiments targeting various domains including color (Bleys, 2008), spatial language (Spranger, 2011), quantifiers (Pauw and Hilfery, 2012) and temporal language (Gerasymova and Spranger, 2012, this volume) on different robotic platforms, including the Sony humanoid (Fujita et al, 2005) and the Humboldt MYON (Hild et al, 2012) robot discussed in earlier chapters of this volume. ...
... More elaborate examples how to map IRL to FCG and syntactic knowledge can be found in (Gerasymova and Spranger, 2012) for temporal language, in (Spranger and Loetzsch, 2011) for spatial language and in (Bleys, 2008) for color. ...
This chapter introduces the computational infrastructure that is used to bridge the gap between results from sensorimotor processing and language. It con-sists of a system called Incremental Recruitment Language (IRL) that is able to configure a network of cognitive operations to achieve a particular communicative goal. IRL contains mechanisms for finding such networks, chunking subnetworks for more efficient later reuse, and completing partial networks (as possibly derived from incomplete or only partially understood sentences).
... Each cognitive operation in IRL takes a number of arguments, one of which (usually called the target argument) is considered to be the result of the cognitive operation. Cognitive operations can be combined together in a network with one operation providing or using results produced by another one, as shown for example in Figure 2 from [31], which represents the procedural semantics of the German phrase "der vordere Block" (the front block). When such a network is executed, each cognitive operation computes specific values for its arguments as fitting with the present world model and discourse context. ...
The paper sketches a methodology for designing and im-plementing complex lexicons and grammars using Fluid Construction Grammar (FCG). FCG emphasizes a functional viewpoint of language and decomposes grammatical systems based on their semantic domains and communicative functions. Rather than directly specifying all the components of a construction explicitly, which would lead to highly com-plex definitions, FCG uses abstractions in the form of templates that implement design patterns common across human languages.
... German spatial language has been used for the reconstruction experiments and the grammatical constructions necessary for expressing spatial relations have been formalized and tested in Fluid Construction Grammar. This grammar is discussed at length in Spranger & Loetzsch (2011). Here we can therefore focus immediately on the acquisition and emergence of spatial language. ...
... As mentioned earlier, we have already conducted reconstruction experiments for German spatial cognition and grammar (Spranger, 2011a;Spranger & Loetzsch, 2011). These experiments used spatial language games within the experimental setup shown in Figure 1. ...
... Hence, when grammar is removed, semantic ambiguity increases. Semantic ambiguity is defined here as the number of different possible semantic structures underlying the same utterance (Spranger & Loetzsch, 2011). ...
This chapter explores a semantics-oriented approach to the origins of syn-tactic structure. It reports on experiments whereby speakers introduce hier-archical constructions and grammatical markers to express which conceptual-ization strategy hearers are supposed to invoke. This grammatical information helps hearers to avoid semantic ambiguity or errors in interpretation. A sim-ulation study is performed for spatial grammar using robotic agents that play language games about objects in their shared world. The chapter uses a re-construction of a fragment of German spatial language to identify the niche of spatial grammar, and then reports on acquisition and formation experiments in which agents seeded with a 'pidgin German' without grammar are made to interact until rudiments of hierarchical structure and grammatical marking emerge.
... The technique is illustrated through German case agreement, whose ambiguity, agreement constraints and feature indetermination are notoriously difficult for the aforementioned traditional solutions. The solution can however easily be applied to other grammatical domains as well, as is shown elsewhere in this book for argument structure (van Trijp, 2011), verbal agreement (Beuls, 2011) and space (Spranger & Loetzsch, 2011). All examples in this paper have been operationalized in Fluid Construction Grammar, but the approach can be implemented in other unification-based formalisms as well. ...
... The same template can be used for any feature that requires an array of values rather than a single value. Other examples in this volume can be found for Hungarian vowel harmony and agreement (Beuls, 2011) and spatial expressions (Spranger & Loetzsch, 2011). Feature matrices have also been applied to German field topology and information structure (Micelli, 2012). ...
This paper illustrates the use of 'feature matrices', a technique for handling ambiguity and feature indeterminacy in feature structure grammars using unification as the single mechanism for processing. Both phenomena involve forms that can be mapped onto multiple, often conflicting values. This paper illustrates their respective challenges through German case agreement, which has become the litmus test for demonstrating how well a grammar formalism deals with multifunctionality. After reviewing two traditional solutions, the paper demonstrates how complex grammatical categories can be represented as feature matrices instead of single-valued features. Feature matrices allow a free flow of constraints on possible feature-values coming from any part of an utterance, and they postpone commitment to any particular value until sufficient constraints have been identified. All examples in this paper are op-erationalized in Fluid Construction Grammar, but the design principle can be extended to other unification-grammars as well.
This paper discusses the application of a procedural semantics framework to autonomous robots. We model insights from decades of research in cognitive linguistics and empirical research to develop a system capable of autonomously producing and interpreting German locative utterances. The system is tested in the real world using a population of robots that talk to each other via natural language.
This chapter compares two computational frameworks developed over the last decade to support investigations into the emergence and use of language, Fluid Construction Grammar (FCG) and Embodied Construction Grammar (ECG). Both of these representational formalisms are rooted in the construction grammar tradition, sharing basic assumptions about the nature of linguistic units and the crucial role played by contextual factors. Nonetheless, they have arisen from different perspectives and with different goals: FCG was designed to support computational language game experiments that address the evolution of communication in populations of robotic agents, while ECG was designed to support cognitive modeling of human language acquisition and use. We investigate how these differing emphases motivated different design choices in the two formalisms and illustrate the linguistic and computational consequences of these choices through a concrete case study. Results of this comparison sharpen issues relevant to computational construction grammar in general and may hold lessons for broader computational investigations into linguistic phenomena.
