Faber, Pamela. 2011. The dynamics of specialized knowledge representation: Simulational reconstruction or the perception–action interface. Terminology 17, no. 1: 9-29.
This is thefinal peer-reviewed versionof the following article: Faber, Pamela. In press. The dynamics of specialized knowledgerepresentation: simulational reconstruction or the perception-action interface. Accepted for publication in Terminology.<http://www.benjamins.com/cgi-bin/t_seriesview.cgi?series=TERM> You can find more articles authored by LexiCon Research Group members at <http://lexicon.ugr.es>. The dynamics of specialized knowledge representation: simulational reconstruction or the perception-action interface PAMELA FABER University of Granada Abstract Dynamicity isthe conditionof being in motion, and thus, is characterized bycontinuous change, activity, or progress. Not surprisingly, dynamicity is generally acknowledged to be an important part of any kind of knowledge representation system or knowledge acquisition scenario. This means that it might be a good idea to reconsider concept representations in Terminology, andmodify them so that they better reflect what the natureof conceptualization inthemind and brain. In this sense, recenttheories ofcognitionhave emphasized thatsituated or grounded experiencesare activated in cognitive processing (Louwerse and Jeuniaux 2010; Barsalou 1999; Zwaan2004).According to these theories, meaning construction heavily relies on perceptually simulating the information that is presented to the comprehender. Specialized knowledge representation that facilitates knowledge acquisition could thus be conceived as asituation model or event that enablescomprehenders to use communicatedinformation to better interact with the world. Keywords:knowledge representation; knowledge acquisition; situated cognition;conceptsystem; terminology resources 1. Introduction Dynamicity is the condition ofbeing in motion, andthus, is characterized by continuous change, activity, or progress. If a frequency check were made in Terminology literature, dynamic would behigh onthekeywordlist. Thisis onlynaturalsincespecializedlanguageis dynamic, and its representation should be so as well. Accordingly, dynamicity is acknowledged to be an important part ofany kindofknowledge representation system or knowledge acquisition scenario. However, an in-depth study isneeded ofthe dynamicity of conceptualization itself, and how the nature of human perception influences the representation of concept systemsin specialized knowledge contexts. As is well-known, a major focus in Terminology and Specialized Communication hasalways beenconceptual organization. In fact, a great deal hasbeen written on the topic Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface (Budin 1994;Puuronen1995;Meyerand Mackintosh 1996;Meyer, Eck, and Skuce 1997; Pozzi 1999;Pilke2001; Feliu 2004; Tebe2005;Faberetal. 2007;León2009, inter alia).Given the fact that terms are specialized knowledge units that designate our conceptualization of objects, qualities, states, and processes in a specialized domain, any theory ofTerminology should aspire to psychological and neurological adequacy. Inthissense, knowledge of conceptualization processes as well as the organization ofsemanticinformation in the brain should underlie any theoretical assumptions concerning the access, retrieval, and acquisition of specialized knowledge as well as the design ofspecialized knowledge resources. It is true that one of the basic premises of Terminology theory from Wüster (1968)onwards is the conceptual organization of terminology resources. Although manyresources do not offer a graphical representation of the conceptual organization of specialized knowledge domains, conceptual information is includedinterm recordsinthe form ofdefinitions and information concerning related terms.Nevertheless, Terminologyhas not as yet seriously taken on board recent research advances in cognition andcognitive neuroscience, which pointto the inadequacyof standard theories of cognition (Gallese and Lakoff 2005). As is well-known, standard theories of cognition are basedon abstract, amodal representations of entities, events, and processes that do not take into account the human and contextual factor of processors, their focus of attention, and spatiotemporal situation.As it happens, these conventional (though inadequate) theories of cognition are the same theories upon which mainstream conceptual representationsinTerminology are currentlybased. This is reflected in Terminology textbooksas wellasinthe design of specialized knowledge resources. Forexample, mostof these manuals mention thefactthat partofterminology work is the elaborationof agraphical representationofaconcept system ofthe specialized field with the help of an expert andthe use of specialized thesauri: A concept system is made up of a structured set of concepts organized intoconceptclasses.Themajorconceptclasses and sub-classes,as well asconcepts of thesame class, are related on the basis of thecharacteristicsthey share orby their use in reality [...] Structures are usually represented in tree diagrams (Cabré 1999: 135). In terminology work, the knowledge acquired in a given subject field isstructured according to the hierarchical and associativerelationships between the concepts that make up the subject field (Pavel and Nolet 2001:15). However, very little is ever said about how this representation is created, and the premises upon which it is based. Various authors have expressed discontent with the current shape ofconcept systems (e.g. Nuopponen 1994; Cabré 2000;Temmerman 2000). Rogers (2004: 221) criticizes the fact that each node in the representation ofaconcept system is conventionally labeled by a decontextualized lexeme despite the fact that knowledge, as represented in texts, is conceptually dynamic and linguistically varied. Quite understandably, dynamicity is difficult to capture and believably portray in a static Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface representation. Perhaps for this reason, there is a certain lack of proposals or viable alternatives to the current state of affairs. The explicit representation of conceptual organization does not appear to have animportantrole in the elaboration of terminologicalresources. Most resources that do offer suchinformation merely provide an overview ofa specialized field, primarily based on the ISA or TYPE_OF conceptual relation. This overview usually consists of graphical representations of concepts in the form of tree orbracket diagrams. However, even this type oforganizationis a fairly rare occurrence since the great majority ofterminologicalresources available on Internet contain very little information regarding the location ofspecialized knowledge concepts in larger knowledge configurations (Faberet al. 2006). Even when conceptual representations are included, they do not correspond to current theoretical accounts of how conceptualization takes place in the mind. Mental representations are much richer and more flexible than such representations ofconceptual structure. Part ofthis perceived richness (as well asthe difficulty in describing it) is due to the inherent dynamicity of conceptual processing and conceptualization,which involves changeover time (Langacker 2001). Because of their dynamic nature, grounded or situated cognition theories are of vital interest for the representation of specialized knowledge, which is a major focus in Terminology and specialized communication. Sinceknowledge resources should reflect, to the extent possible, conceptual categories and the processes that actually occur in thebrain,itis timethat terminologists took note ofrecent advances in cognition, and made aneffort to model specialized knowledge representations accordingly. The question is howan awarenessof the natureof mentalprocessescanbe appliedto and incorporated in the representation of specialized knowledge concepts. 2. New theories of cognition Recent research in cognitive psychology and neuroscience highlights the dynamic nature of categorization, conceptstorage and retrieval, andcognitiveprocessing (Louwerse and Jeuniaux2010; Aziz-Zadeh and Damasio 2008; Patterson, Nestor and Rogers 2007; GalleseandLakoff 2005).This workunderlinestheinadequacy ofstandard theories of cognitionthat claim that knowledge resides in asemantic memory system separate from thebrain’smodal systems forperception, action, and introspection. According to standard theories, representations in modal systems are not greatly influenced by the perceiver and thecontext of perception, and are transduced into amodal symbols, which are not specific of the mode of perception. These symbols represent knowledge about experience in semantic memory (Barsalou 2008:618; Mahon and Caramazza2008:59). However, there isan increasingconsensus in favor ofamore dynamic view of cognitiveprocessing or situated cognition, whichreflects the assumptionthat cognition is typically grounded in multiple ways. These include simulations, situated action, and even bodily states. The embodied or grounded cognition hypothesis equates understanding with Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface sensory and motor simulation. This hypothesis claims that interactions betweensensorimotor systems and the physical world underlie cognition. When we encounter aphysical object, our senses represent it duringperception and action. Processing the objectinvolves partially capturing property information on these modalities so that thisinformation can later bereactivated (Damasio andDamasio1994). For example,to representthe concept, PEACH, neural systems for vision, action, touch, taste and emotion partially reenact theperceiver’s experience of apeach. These reenactments or simulations are not the samethingasmentalimagery, whichis consciously evoked inworking memory. Unlike mental imagery, these simulations seem to be relatively automatic processes that lie outside of our awareness (Simmons, Martin and Barsalou 2005: 1602). To date, brain-imagingexperiments have largely involved everyday objects suchas cups,hammers, pencils, and food, which, when perceived, trigger simulations of potentialactions.Forexample, the handleof acup activates a graspingsimulation (Tucker and Ellis 1998, 2001).Food activates brain areas related togustatoryprocessing aswellas areasinthe visual cortex representing object shape (Simmons, Martin and Barsalou 2005). Neuroimaging research thus confirms that simulation is a key part of conceptualprocessing (Martin 2001, 2007). When conceptual knowledge about objects isrepresented, brain areas that represent their properties during perception and actionbecome active.Inparticular,brainareasthat represent the shape and color of objects, the motion they exhibit, and the actions that agents perform on them become active to represent these properties conceptually. Such reenactments not onlyoccurin the presenceof the objectitself, but also in response to words and other symbols.It would thusappear thatsimulations have a centralroleinthe representation of conceptual knowledge (Barsalou 2003; Martin 2001, 2007). For precisely this reason, they should betaken into account in Terminology. Tomy knowledge,few if any neuropsychological experiments of this type have ever been performed withspecialized concepts, but there is no reason to suppose that the brain would work any differently. For example, when reading about hockey, experts were found to produce motor simulations absent in novices (Holt and Beilock 2006). In all likelihood, a similar result would be obtained if the object were a tide gauge, pluviometer, or anemometer. The expert’s brain would showmotorsimulations inbrain areas that would not beactivatedinthe case of non-experts to whom the object was unfamiliar. The information regarding simulated interaction is thus a vital part of conceptual meaning. The nature ofsuch simulations is componentialrather thanholistic.In other words,theyare not continuous streamed video recordings, but rather contain many small elements ofperception, which arise from all modalities of experience. These modalities are contextually constrained and vary in accessibility (Simmons, Martin and Barsalou 2005). This would(or should) haveasignificanteffect onhowconcepts are defined and how the Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface definition is structured. The way that objects are represented in our brain seems to suggest that current methods and ways of elaborating specialized knowledge representations should be modified in order totake this information into account. 3. The dynamics of Terminology Yet, we may well ask ourselves ifsuch research on the dynamicity ofcognition, howevervaluable, has any feasible application in Terminology and specialized communication.Terminological dynamicity has been explored from a wide variety of perspectives. For example, our knowledgeof specializedfieldsevolves, and the terms usedto describe the conceptsin them also change (BowkerandPearson2002: 48). Dynamicity isaproperty of termformationas exploredin Kageura(2002).Itunderlies theidea oftheemergence ofterms, the coherent coming into existence of new forms through ongoing intrinsicprocesses. Dynamism is also reflected in the historical evolution of term meaning withinsociocultural context (e.g. splicing, Temmerman 1995, 2007). Moreover, the dynamic nature ofterms and theirconstantchange in meanings may require humanintervention in the formofterminological control(Felber 1988; Oeser and Budin 1995). However, what underlies allof these dynamic perspectives is the fact that conceptualization or concept formation itself is dynamic. This is the process through which we access and acquire knowledge. In reference to dynamic conceptualization, Wright (2003) and Antiaet al. (2005) referto Damasio (1994) and the dynamic variability of his earliest model of concept formation. The model of memory described is reconstructive. Concepts take the form of fleetingperceptions, which are essentially instantaneous convergencesofperceptual aspects that combineat a given point intime and space.Themain conclusion seems tobe thatconcepts stemfroma series of iterativeprocessing events, and are inconstant flux in the brain. Suchextreme dynamism initially seems to have a very limited practical application because itisimpossible to capture a process that is in perpetual motion and so individual. However, the position that semantic memory arises from universal connectivity in the brain without a corresponding stable neuralarchitectureis no longer tenable (Patterson,Nestor and Rogers 2007:976). Although current theoretical positions regarding semantic memory share the view that much of our semantic memory relates to perception and action, in order to be able to generalize across concepts that have similar semanticsignificance, there must also be a single convergence zone or hub that supports the interactive activation of representations in all modalities for semantic categories (Patterson, Nestor and Rogers 2007: 977). The question is how such theories can affect or be applied to terminological work, including the creation of terminological resources. Actually, theyhave a range ofpossibleapplications in Terminology that are just beginning to be explored. First of all, situated conceptualizations underline thefact that concepts are notprocessed in isolation, but aretypically situated in background situations and events (Barsalou 2003). This signifies that context is all-important in knowledge representation. At any given moment in the Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface perception of the entity, people also perceive the space surrounding it, including the agents, objects, and event present in it (Barsalou 2009: 1283). This can be directly applied to specialized knowledge modeling to ensure the comprehensiveness of terminological entries. In fact, it can act as a safeguard against omitting other closely related concepts in the same knowledge domain. For example, EROSION is the wearing away of the earth’s surface, but whether conceptualized as aprocess or the result of this process, erosion cannot be conceivedinisolation.Itisinduced by anagent(wind, water,orice)that affects a geographicentity(the Earth’s surface)bycausing something (solids) to move away. Moreover, any process takes place over aperiod oftime, and can bedividedinto smaller segments.Inthis sense, erosion can happen at aspecific season of the year, and may takeplacein acertain direction. All of this information abouterosion should be available forpotentialactivationwhen the userwishes toacquireknowledge about it. The meaning of aconcept is constructed on-line, and is modulated by context. Secondly, althoughdynamicity has been regarded primarily as an attributeofevent and action concepts (Pilke 2001; Puuronen 1995, inter alia), as shall be seen, grounded orsituated cognition means that objectconcepts are also dynamic since they are processedas part of a frame or dynamic context which highlights the type of action that they participate in. This, in turn, affects how concepts should be represented in order to facilitate knowledge acquisition and understanding. Thirdly, research results in this area indicate that knowledge acquisition requires simulation of human interaction with objects, and this signifies that horizontal or non-hierarchical relations that define thegoal, intended purpose, affordances, and resultof the manipulationand use ofan object (e.g. HAS_FUNCTION,AFFECTS, HAS_RESULT, etc.)are just as important as vertical or hierarchical ones, such asTYPE_OForPART_OF. 4. Frame-based Terminology and dynamic knowledge representation Simulation represents the way we interact with an entityand how entities interact witheach other. This means that no specialized knowledge concept can be activated inisolation, but rather as part of an event. When this is applied to Terminology and specialized communication, this has the effect of making context or situation a crucialfactor in knowledge representation. Our knowledge of a concept initially provides the context or event in which it becomes meaningful for us. A knowledge resource thatfacilitatesknowledge acquisition should thus provide conceptualcontexts or situations inwhich a concept is related to others in a dynamic structure that can streamline the action-environment interface. Rather than being decontextualized and stable, conceptual representations should be dynamically contextualized to support diverse courses ofgoalpursuit (Barsalou 2005: 628). Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface Frame-based terminology (Faber, Marquez and Vega 2005; Faber etal2006; Faber, Leon, Prieto and Reimerink2007) uses a modified version ofFillmore’s Frames (Fillmore 1982, 1985; Fillmore and Atkins 1992) coupled with premises from Cognitive Linguistics to configure specialized domains on the basis of definitional templates and create situated representations for specialized knowledge concepts. The compatibility of CognitiveSemantics with neuropsychological research on category-specific semantic deficits isunderlined in Rodriguez (2004). 4.1. Event representation In Frame-based Terminology, conceptual networks are based on an underlying domain event as well as a closed inventory of both hierarchical and non-hierarchical semantic relations. We have used these premises to construct an environmental knowledge base called EcoLexicon (http://ecolexicon.ugr.es). The mainfocus ison conceptual relationsas well as a concept’s combinatorial potential, extracted from corpus analysis. This prototypical domain event or action-environment interface (Barsalou 2003) provides a template applicable to all levels of information structuring. Figure 1. Environmental Event In EcoLexicon, knowledge can be accessed from top-level categories to more specificrelational structures. The most generic level is the Environmental Event (EE), whichprovides aframe for theorganizationof all concepts in the knowledgebase.As showninFigure 1, the EE is conceptualized as adynamic process that is initiated byan agent(either naturalor human). This process affects a patient (an environmentalentity), and produces Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface a result. These categories(agent, process, patient,etc.)arethe concept roles characteristic of this specialized domain. Additionally, there are peripheral categories which includeinstruments that are typically used during the EE, as well asa categorywheretheconcepts of measurement, analysis, and description of the processes in the main event are included. This event-based representationfacilitates knowledge acquisition intext processing since conceptual categories arebound together byeventknowledge. Proof ofthe usefulness of event knowledge canbe foundinwritten communication since a comprehender’s knowledge of events plays a central role in sentence processing. Thisknowledge interacts with structural interpretation at the earliest possible moment (Elman2009: 549). Evidently, terms, whether they designate objects orprocesses, are powerful cuesforthewider event knowledge targeted. Inthis regard, the choice ofa specific term is enough to generate expectations and predictions that constrain the range oflikely events. 4.1.1. EXTREME EVENT For example, one of the concepts in EcoLexicon is EXTREME EVENT in its sense ofnaturaldisaster.Disastersin the environmentinclude great earthquakes,floods, giant sea waves, hurricanes, tornadoes, etc., and their consequences. The concept of EXTREME EVENT isvery complex since it is a natural agent that initiates a process (i.e. earthquakes or volcanic eruptions can produce tsunamis) butitcanalso betheprocessitself, which occurs in time and space. This information is represented in EcoLexicon as shown inFigure 2. Figure 2. Representation ofEXTREME EVENTin EcoLexicon Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface As showninFigure 2,allof the conceptsclosest tothecentralconcept are connectedtoit bya seriesof conceptualrelations thatareexplicitly named(e.g. TYPE-OF, CAUSES,AFFECTS,etc.). These conceptual relations are the graphical representation of the informationimplicit in the definition, which appears when thecursor is placed on the main concept. Since EXTREME EVENT is a very general concept, the only visual information that can be associated with it is that ofits subtypes (HURRICANE, TORNADO,EARTHQUAKE,FLOOD, etc.). The majority ofrelations at this level are thus TYPE_OF. However, EXTREME EVENT also activates non-hierarchical relations typical of the general event frame. As such, its principalattribute is RISK;itAFFECTS the environment; and CAUSES an environmental impact. As forthe TYPE_OF relations, they can be regarded as access routes to more prototypical base-level concepts (Rosch 1978), which do have a mental image, and can activate specific contexts. This set of subtypes (hurricane, tornado, flood, tsunami, etc.) takethe form of constellations, eachwiththeir own set ofsubordinateconceptsand conceptual relations, which encode more specific sub-event knowledge and representations. 4.1.2. Recontextualization: HURRICANE According to Barsalou (2005), a given concept produces many different situatedconceptualizations, each tailored to different instancesin different settings.Thus, contextcan be said to be a dynamic construct that triggers or restricts knowledge. This generalevent that codifies a natural disaster can thus be recontextualized at any moment to centeron any of the more specific subevents. For example, when the EXTREME EVENT representation is recontextualized to focus onHURRICANE,it takes the followingform. Figure 3. Representation ofHURRICANE in Ecolexicon Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface This type of recontextualization of EXTREME EVENT still contains asector of the previousinformation, but varies the focus of attention so that hurricaneis now thecenteroffocus. Besides communicating the fact that hurricane is a type of extreme event, this new representation highlights the fact that wind and flooding are crucial participants in the event. Wind is part of ahurricane, and a hurricane causes floods. Not surprisingly, WINDand FLOOD are concepts that are susceptible to simulation since they can directly affect humanlife and health.Italso mentionstheattributeoflow atmospheric pressure as wellas the scale used for hurricane measurement (Saffir-Simpson hurricane scale), which codifies an important aspect of expert interaction with a hurricane. 4.2. Object representation Object concepts canalsobe represented dynamically aspartsofevents. They arestored in semantic memory, a major division of declarative memory, which contains informationaboutthemeaningof objects and words.Thisisthepart ofour mind(oratleasta small sectionofit) thatterminologists are tryingtomodel eachtimethey try tomakea concept map. Howknowledge is modeled largely dependson how objects are defined, their focalproperties, their perceived relations with other concepts, andhow the user understandsthem. Accordingly, one of the basic characteristics underlying the representation of objects is knowledgeof whether theycanbe manipulated andifso,exactly how thisis done.Inthe case ofman-made objects, another important property is their function, or how they can be used. Thiswouldmean thatanimportant partof theinformation in the representation of specialized engineering instruments would evidently involve how they are used byhumans, for what purpose, and what is the result of themanipulation. 4.2.1. RECORDING INSTRUMENT For example, aRECORDING INSTRUMENT(e.g.marigraph, pluviograph, anemograph, etc.)is a subtype of INSTRUMENT. As aman-made manipulable artifact,a recording instrument has a function (i.e.recording) as well as an object that is recorded(tides, rain,wind, etc.).As a tool, it is strongly susceptible to human interaction, and activates a simulation frame in which much ofthe perceiver’s knowledge of the artifact involves his/her ability to handle it and insome way toextract information from it. For example, Figure 4 shows the representation ofPLUVIOGRAPH. The representation of PLUVIOGRAPH,ofcourse, includesTYPE_OFinformation.Apluviograph is a recording instrument, and has subtypes, such as digital pluviograph and portable pluviograph. However,itisalso partof whatmightbe calledaRECORDING EVENT inwhich a humanagentcauses the machine to record and generate a representation of something (RAINFALL).The recording instrument used in thiseventis a pluviograph,which produces (or effects) aPLUVIOGRAM. As can beobservedinFigure 4,this processis reflectedinthe non-hierarchical relations REPRESENTSandEFFECTED_BY. Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface Figure 4. Representation ofPLUVIOGRAPH in EcoLexicon 4.3. Domain-specific representation Recent research in cognitive neuroscience also has implications for specialized domains and their organization. Whether conceived as a conceptual category (e.g. GEOGRAPHIC OBJECT, STORM-EVENT, MARITIME CONSTRUCTION,etc.) or as a specialized field of knowledge (e.g.GEOLOGY,ENGINEERING,etc.), domains and domain structure are centralto any theory ofTerminology and specialized communication. Not surprisingly, domains have also been found to exist in the brain in some form, as shown in the large body of research oncategory-specific semantic deficits (Warrington and McCarthy 1983, 1987; Warrington and Shallice 1984; Humphreys and Forde 2001; Caramazza and Mahon 2003; Martin2007; Mahon and Caramazza2008, 2009, inter alia). Although initially research did not provide conclusive evidenceof the important role ofcategories, the domain-specific hypothesis (Caramazza and Shelton 1998) assumes that the first-order constrainton the organizationofinformation within the conceptual system or the organizationof conceptualknowledgein the brainisobject domain. In this model,object domain and sensory, motor, and emotional properties jointly constrain the organization of conceptual knowledge. In addition, object domain is a first-orderconstraint on the organization of informationat both aconceptual level as well as at the Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface level of modality-specificvisualinput representations (Mahon and Caramazza 2009: 34). Although Mahon and Caramazza (2009: 30) restrict basic domains to those with an evolutionary relevanthistory (e.g. living animate, living inanimate, conspecifics, and tools),their observation that domains are constrained by the nature of concept members has evidentimplications for Terminology. One conclusion that can be derived from this hypothesis is the fact that not all categories are structured in the same way, andthat organization is constrained in some significant wayby the natureof the categoryitself.Accordingly, the analysis ofthe properties shared by category members provides a general representational template for each category, which makes the definition of category members more systematic.In this way, definitions acquire a more uniform structure that complements the information encoded in the conceptual system, and directlyrefers toandevokes the underlying event structure ofthe domain. Such templates can even be considered a kind of conceptual grammar (Faber, Leon, Prieto and Reimerink 2007). However, category templates are madeup ofdifferent clusters of conceptual relations that dependon thenature of the category. 4.3.1. Domains as conceptual categories In Terminology, there are two different ways ofconceivingspecialized domains. Domains can either be viewed as conceptual categories or as specializedknowledge fields. Whendomains areconceptual categories, categories are constrained by the nature of category members that share properties. For example, the categories of specialized instrumentsand geographic objects are quitedifferent from each other. This can be observed inthe conceptual relations thatreflect their interconnections with other entities. As previously mentioned,the INSTRUMENT domain isprimarily constrained byinformation regarding properties related to manipulation, function, and result. This is invivid contrast to a domain such as GEOGRAPHIC OBJECTS (estuary, marshland, channel, etc.) which is constrainedby other types ofinformation, directly linked to the nature of the concepts.ASSmith and Mark (1999)point out,the specificities of geographic objects are the following: . Geographic objects areintrinsically tied to their location in space [LOCATION_OF]. . Theyare often size- or scale-dependent [SIZE_OF]. . They are often the products of delineation within a continuum in which other objects, including human agents, liveand move[DELIMITED_BY]. This cluster of relations stem from the fact that geographic objects are presumably simulatedin a differentwayfrom instruments, atmospheric phenomena, coastal defense structures, or marine fauna, and this affects their conceptualization and representation. Although this would require further study, the simulation of geographic objects wouldinvolve the activation of brain areas connected with location and orientation. For this reason, within this category, emphasis has been placed on information pertaining to spatial orientation. Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface Figure 5. Representation ofMARSHLAND in EcoLexicon As shown inFigure 5, the representation ofMARSHLAND ENVIRONMENTactivatesa different set of relations from event and disaster concepts. As a geographic concept, MARSHLAND ENVIRONMENTis represented as beingdelimited by the sea or ariver. It isan environmental area affected by floods. Lagoons, tidal flats and marshes are also geographic objectslocated in MARSHLAND ENVIRONMENT.This is indicative ofits ample size, which means that itcan include a wide variety of geographic concepts. 4.3.2. Domains as specialized knowledge fields As has been shown, concepts within a domain areinternally constrained by the nature of the domain. When domains are conceived of as specialized knowledge fields, such asChemistry, Geology or Engineering, they provide further contexts in which versatile concepts are recontextualized(Leon,Magaña, and Faber 2009; Leonand Magaña, 2010). This is another source of dynamicity. For example,even thoughWATERis not a specialized conceptper se,it mustbeincludedin anyenvironmentalknowledgebase sinceitis central tomany specialized environmental processes and object representations. Given that WATERis a concept that participates in so Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface manyother representations ofenvironmental concepts, the conceptual information linked to it must be contextualized so as not to generate an information overload. This means that contextual constraints must be applied so as only to activate the conceptual relationsrelevant to WATERin a given specialized field (Leon, Magaña and Faber 2009; Leon and Magaña 2010).This istheonlywayto eliminate undesired information. In Figure 6, the representation ofWATERemphasizes the affordances that water has for engineering. Figure 6. WATER in the context of Engineering More specifically, it highlights the fact that water is a building material for engineering structures, and is used in processes such as pumping and dredging. The most salient conceptual relations are thus MADE_OF and AFFECTS. In contrast, when water isrecontextualized within the geology context, its representation is quitedifferent since inthis context,information regarding how waterinteracts withsoil andlandscapeismuch more important. Consequently, another set of concepts and relations are activated. Evidently, the number of conceptual relations varies from one network to another. Relation types also differ, which highlights the changing nature of water’s internalstructure according to each semantic role. For example, in Engineering, most relations to WATER are MADE_OF and AFFECTS, whereas in Geology, CAUSES and TYPE_OF are the most salient conceptual relations (Leon andFaber 2010). In this sense, in EcoLexicon, we are currently exploring the possibility of establishingcontextual field-related constraints onthe activation ofconceptual relations. This would beappliedto general objects and processes, suchasWATER, OCEAN,SEDIMENTATION, EROSION,etc., which otherwise would generate an excess of information. Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface Figure 7. WATER in the context ofGeology Nevertheless, recontextualization does not involve a clear-cut distinction between different context domains since theycanalso share certain conceptual propositions. This is due tothefact thatmultidisciplinarity gives rise to fuzzy categoryboundaries and asa result, contextual domains can form their own hierarchical structure. Moreover, they are also dynamic flexiblestructures thatshould evolveover time according to the type andamount of informationstored in our knowledge base (León andMagaña 2010). 4. Conclusion Dynamicity is a crucial issue in Terminology because it lies at the root of specialized communication and knowledge representation. However, it is rarely adequately reflected in terminological resources. Reasons for this include the difficulty of portraying dynamic processes by means of conceptual trees. Such representations stem from standard theories ofcognition, based on the abstract, amodal representation of entities, events, and processes. However,amore dynamic viewof cognition, derived from recent research inneuroscience, claims that understanding islargely based on sensory and motor simulationwith possibly a single convergencezone that affords the possibility to generalize across Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface concepts that have similar semantic significance. This has evident applications to Terminology and its dynamic nature, which include the following: 1. No specialized knowledge concept should be activated in isolation, but ratheras part of a larger structure or event. 2. A specialized knowledge resource that facilitates knowledge acquisition shouldthus provide conceptual contexts or situations in which a concept is related to others in a dynamic structure that can streamline the action-environmentinterface. Within this context, all concept types are regarded as dynamic because they are partof a process or event. 3. Since knowledge acquisition and understanding requires simulation, this signifiesthat non-hierarchical relations defining goal, purpose, affordance, and result of themanipulation and use of an object are just as important as hierarchical generic-specific and part-whole relations. 4. Research proposals, such as the domain-specific hypothesis (Caramazza and Shelton 1998) also have implications for Terminology since it asserts that domains areconstrainedby the natureoftheir members.In Terminology, thisis reflectedinclusters of conceptual relations that make up the general representationaltemplate, characterizing different categories. All of these conclusions have been illustrated by examples from EcoLexicon, an environmental knowledge base (available at: http://ecolexicon.ugr.es). EcoLexicon is aconceptually-organized, frame-based terminological resource that facilitates knowledge acquisition since it presents concepts as part oflarger knowledge structures and permitsdynamicprocesses such as the recontextualizationof knowledge representations. They also point to the fact that the representation of specialized knowledge concepts should incorporate dynamicity at some level. Only in this way will terminological resources be more effective and facilitate knowledge acquisition. Acknowledgements This research was funded by the Spanish Ministry of Science and Innovation (project FFI2008-06080-C03-01/FILO). References Antia, B, E., Budin, G., Picht, P., Rogers, M., Schmitz, K. D. and Wright, S.E. 2005. Shaping translation: a view from terminology research. META 50 (4). Available at: http://id.erudit.org/iderudit/019907arAziz-Zadeh, L. and A. Damasio. 2008. “Embodied semantics for actions: Findings from functional brain imaging.” Journal of Physiology – Paris 102, 35–39. Barsalou, L. W. 1999. “Perceptual symbol systems.” Behavior and Brain Sciences 22, 577– 660. Barsalou L. W. 2003. “Situated simulation in the humanconceptual system.”Language and Cognitive Processes 18, 513–62. Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface Barsalou, L. W. 2005. “Situated conceptualization. In Cohen, H. and Lefebvre, C. (eds.), Handbook of Categorization in Cognitive Science. St. Louis: Elsevier. Barsalou, L. W. 2008. “Grounded cognition.”Annual Review of Psychology 59, 617–645. Barsalou, L. W. 2009. “Simulation, situated conceptualization, and prediction.” Philosophical Transactions of the Royal Society B, 1281–1289. Barsalou, L. W., C. Breazeal, and L. B. Smith. 2007. “Cognition as coordinated non-cognition.”Cognitive Processing 8(2), 79–91. Bowker,L. andJ.Pearson. 2002.Working with Specialized Language. A Practical Guide to Using Corpora. London/New York: Routledge. Budin, G. 1994. “Some hypotheses about concept representations.”InProceedings of the 9th European Symposium on LSP, Bergen, Norway, 2-6 August, 1993. Bergen: Fagbokforlaget. Cabré, M. T. 1999. Terminology Theory, Methods and Applications. Amsterdam/ Philadelphia: John Benjamins. Cabré, M. T. 2000.“Theories of Terminology.” Terminology 6(1), 35–57. Caramazza, A. and B. Z. Mahon. 2003. “The organization of conceptual knowledge: the evidence ofcategory-specific semantic deficits.” Trends in Cognitive Sciences 7(8), 354– 361. Caramazza, A. and J. R. Shelton. 1998. “Domain specific knowledge systems inthebrain:the animate-inanimate distinction.” Journal of Cognitive Neuroscience 10, 1–34. Damasio, A. 1994. Descartes’ Error: Emotion, Reason, and the Human Brain. New York: Avon. Damasio, A.andH. Damasio.1994.“Cortical systemsfor retrieval ofconcrete knowledge: the convergence zone framework.”In Koch, C. andJ. Davis (eds.), Largescale Neuronal Theories of the Brain. Cambridge, MA: MIT Press. Elman,J.L. 2009.“On themeaning words and dinosaurbones: lexical knowledge without a lexicon.”Cognitive Science 33, 547-582. Faber, P., P. León, J. A. Prieto, and A. Reimerink. 2007. “Linking images and words: the description of specialized concepts.” International Journal of Lexicography 20, 39–65. Faber, P., C. Márquez, and M. Vega 2005. “Framing Terminology: A Process-Oriented Approach.”META 50(4). Faber, P.,S.Montero, M.R. Castro,J.Senso,J.A. Prieto, P. León, C.Márquez, and M.Vega. 2006. “Process-oriented terminology management in the domain of CoastalEngineering.”Terminology 12(2), 189–213. Felber, H. 1988. Korollierte Begriffsdynamik. Berlin: Cedefop. Feliu, J. 2004. Relacions conceptuals I terminologia: analisi i proposta de deteccio semiautomatica. PhD thesis, Barcelona: Universidad Pompeu Fabra, Instituto Universitario de Linguistica Aplicada (IULA). Fillmore, C.J. 1982.“Framesemantics.”In Linguistics Society ofKorea(ed.),Linguistics in the Morning Calm. Seoul: Hanshin. Fillmore, C.J. 1985.“Frames and the semantics ofunderstanding.” Quaderni di Semántica 6(2), 222–254. Fillmore, C.J. and B.T. SAtkins. 1992.“Towardsaframe-basedlexicon: the semantics of risk anditsneighbours.”In Lehrer, A.andE.Kittay(eds.),Frames, Fields and Contrasts, Hillsdale, NJ: Lawrence Erlbaum. Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface Gallese, V. andG. Lakoff. 2005. “The brain’s concepts:the role ofthe sensory-motor system in conceptual knowledge.” Cognitive Neuropsychology 22 (3/4), 455–479. Glenberg, A. M. 1997. “What memory is for.” Behavioral and Brain Sciences 20, 1–55. Holt L.E. and S.L. Beilock.2006.“Expertiseanditsembodiment:examiningtheimpactofsensorimotorskill expertise on the representationofaction-related text.” Psychonomic Bulletin and Review 13, 694–701. Humphreys, G. W. and E. M. Forde. 2001. “Hierarchies, similarity, and interactivity inobject recognition: ‘category specific’ neuropsychological deficits.” Behavioral and Brain Sciences 24, 453–4509. Kageura, K. 2002. The Dynamics of Terminology. Amsterdam/Philadelphia: John Benjamins. Langacker, R. 2001. “Dynamicity in grammar.” Axiomathes 12,7–33. León, P. 2009. Representación Multidimensional de Conocimiento Especializado. PhD thesis. Granada: University of Granada. León, P. and P. Faber. 2010. “Natural and contextual constraints for domain-specificrelations.” In Proceedings of LREC 2010 (Workshop: Semantic Relations. Theory andApplications), 18–21May 2010, Valetta, Malta. Leon, P. and P. Magaña.2010.“EcoLexicon: contextualizingan environmental ontology.” In Proceedings of the Terminology and Knowledge Engineering Conference (TKE), 12–13 August, 2010, Dublin, Ireland. León, P.,P.Magaña, and P.Faber. 2009.“Building the SISE:anenvironmental ontology.” InHrebícek, J., J. Hradec, E. Pelikán, O. Mírovský, W. Pilmmann, I. Holoubek, and R. Legat. (eds.), Towards eEnvironment (Challenges of SEIS and SISE: Integrating Environmental Knowledge in Europe). Available at: http:/www.e-envi2009.org/proceedings/. Louwerse,M. M.andP.Jeuniaux. 2010.“Thelinguistic and embodied natureof conceptualprocessing.”Cognition 114, 96–104. Mahon, M. Z. and A. Caramazza. 2008. “A critical look at the embodied cognition hypothesis and a new proposal for grounding conceptual content.” Journal of Physiology–Paris 102, 59–70. Mahon, M. Z. and A. Caramazza. 2009. “Concepts and categories: A cognitive neuropsychological perspective.” Annual Review of Psychology 60, 27–51. Martin A. 2001. Functional neuroimaging of semantic memory. In Cabeza, R. and A. Kingstone (eds.), Handbook of Functional NeuroImaging of Cognition, Cambridge, MA: MIT Press. Martin,A. 2007. “Therepresentation ofobject concepts inthe brain.” Annual Review of Psychology 58, 25–45. Meyer, I. and K. Mackintosh 1996. “Refining the terminographer's concept-analysis methods: How can phraseology help?”Terminology 3 (1), 1–26. Meyer, I., K. Eck, and D. Skuce. 1997. “Systematic concept analysis within a knowledge-based approach to terminology.” In Wright, S. E. and G. Budin (eds.), Handbook of Terminology Management. Amsterdam/Philadelphia: John Benjamins. Nuopponen, A. 1994. “Causal relations in terminological knowledge representation.” Terminology Science and Research 5(1), 36–44. Oeser, E.andG. Budin 1995.“Controlledconceptual dynamics:from ordinary language to scientific terminology –and back.” Terminology Science and Research 6 (2), 3–17. Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface Patterson, K., P. J. Nestor, and T.T. Rogers. 2007. “Wheredoyou know what you know? The representation of semantic knowledge in the human brain.” Nature Reviews Neuroscience 8, 976–988. Pavel, S.andD. Nolet.2001. Handbook of Terminology. Canada: Minister of Public Works and Government Services. Pilke, N, 2001. “Field-specific features of dynamic concepts – What, when and why?” InMayer, F. (ed.), Language for Special Purposes: Perspective for the New Millennium.Tübingen: Gunter Narr. Pozzi, M. 1999. “The concept of ‘concept’ inTerminology: aneed for a newapproach.” InSandrini, P. (ed.). TKE’99 Terminology and Knowledge Engineering Proceedings, Fifth International Congress on Terminology and Knowledge Engineering, 23–27 August, 1999. Vienna:TermNet. Puuronen, N,1995.“On describing dynamic concepts– A philosophical and terminological approach.” In Budin, G. (ed.), ITTF Proceedings of the 10th European LSP Symposium.Vienna:TermNet. Rodriguez, A. L. 2004. “Aspects of cognitive linguistics and neurolinguistics: conceptual structure and category-specific semantic deficits.” Estudios Ingleses de la Universidad Complutense 12, 43–62. Rogers,M. 2004. “Multidimensionality in concepts systems.” Terminology 10 (2), 215 – 240. Rosch, E.,1978.“Principles ofcategorization.”In Rosch,E. andB.B. Lloyd(eds.),Cognition and Categorization. Hillsdale, NJ: Erlbaum.Simmons, W.K., A. Martin, and LW. Barsalou. 2005. “Pictures ofappetizingfoods activate gustatory cortices for taste and reward.” Cerebral Cortex 15, 1602–1608. Smith, B. and D. Mark. 1999. “Ontology with human subjects testing: An empirical investigation of geographic categories.” American Journal of Economics and Sociology 582, 245–272. Tebe, C. 2005. La representacio conceptual en terminologia: l’atribucio tematica en els bancs de dades terminologiques. PhD thesis. Barcelona: Universidad Pompeu Fabra, InstitutoUniversitario de Lingüística Aplicada (IULA). Temmerman,R. 1995. “The process ofrevitalisation ofold words: ‘Splicing’, a case study in the extension of reference.” Terminology 2(1), 107–128. Temmerman, R. 2000. Towards new ways of terminology description. The sociocognitive approach. Amsterdam/Philadelphia:John Benjamins. Temmerman,R. 2007 “Sociocultural situatednessof terminology in the life sciences: The history of splicing.” In Frank, R., R. Dirven, J. Zlatev, and T. Ziemke (eds.),Language and Mind. Vol II. Interrelations between Biology, Linguistics and Culture. Berlin: Springer. Tucker,M. and R.Ellis. 1998.“On the relations between seenobjects and components ofpotential actions.” Journal of Experimental Psychology: Human Perception and Performance 24, 830–46. Tucker, M. and R. Ellis. 2001. “The potentiation of grasp types during visual object categorization.”Visual Cognition 8, 769–800. Warrington, E. K. and R. McCarthy. 1983. “Category specific access dysphasia.” Brain 106, 859–878. Faber (2011) The dynamics of specialized knowledge representation: simulational reconstruction or the perceptionaction interface Warrington, E. K. and R. McCarthy. 1987. “Categories of knowledge: further fractionationsand an attempted integration.” Brain 110, 1273–1296. Warrington, E. K. and T. Shallice. 1984. “Category-specific semantic impairment.” Brain107, 829–854. Wright, S. E. 2003. From the semiotic triangle to the semantic web. Journal of the International Institute for Terminology Research 14, 111–135. Wüster, E. 1968. The Machine Tool; an Interlingual Dictionary of Basic Concepts. London:Technical Press. Zwaan, R.A. 2004. “The immersed experience: Towardan embodiedtheory of language comprehension.” In Ross,B. H. (ed.),The Psychology of Language and Motivation, New York: Academic Press.
© LexiCon Research Group