Cogn Process DOI 10.1007/s10339-013-0594-9
RESEARCH REPORT
The embodied nature of medical concepts: image schemas and language for PAIN Juan Antonio Prieto Velasco Maribel Tercedor Sa´nchez
•
Received: 15 July 2013 / Accepted: 12 December 2013 Marta Olivetti Belardinelli and Springer-Verlag Berlin Heidelberg 2014
Abstract Cognitive linguistics assumes that knowledge is both embodied and situated as far as it is acquired through our bodily interaction with the world in a specific environment (e.g. Barsalou in Lang Cogn Process 18:513–562, 2003; Connell et al. in PLoS One 7:3, 2012). Therefore, embodiment provides an explanation to the mental representation and linguistic expression of concepts. Among the first, we find multimodal conceptual structures, like image schemas, which are schematic representations of embodied experiences resulting from our conceptualization of the surrounding environment (Tercedor Sa´nchez et al. in J Spec Transl 18:187–205, 2012). Furthermore, the way we interact with the environment and its objects is dynamic and configures how we refer to concepts both by means of images and lexicalizations. In this article, we investigate how image schemas underlie verbal and visual representations. They both evoke concepts based on exteroception, interoception and proprioception which can be lexicalized through language. More specifically, we study (1) a multimodal corpus of medical texts to examine how image schemas lexicalize in the language of medicine to represent specialized concepts and (2) medical pictures to explore the depiction of image-schematic concepts, in order to account This article is part of the Special Section on ‘‘Embodied Social Cognition,’’ guest-edited by Fernando Marmolejo Ramos and Amedeo Dangiulli. J. A. Prieto Velasco (&) Department of Philology and Translation, University Pablo de Olavide, Ctra. de Utrera km. 1, 41013 Seville, Spain e-mail: [email protected] M. Tercedor Sa´nchez Department of Translation and Interpreting, University of Granada, C/Buensuceso, 11, 18002 Granada, Spain e-mail: [email protected]
for the verbal and visual representation of embodied concepts. We explore the concept PAIN, a sensory and emotional experience associated with actual or potential tissue damage, using corpus analysis tools (Sketch Engine) to extract information about the lexicalization of underlying image schemas in definitions and defining contexts. Then, we use the image schemas behind medical concepts to consistently select images which depict our experience of pain and the way we understand it. Finally, such lexicalizations and visualizations will help us assess how we refer to PAIN both verbally and visually. Keywords Image schemas Embodiment Knowledge visualization Medical concepts
Introduction Research on cognition and language understanding has paid close attention to the discussion about whether conceptual processing is primarily embodied or linguistic. On the one hand, there is experimental evidence for the existence of a major embodied component (Barsalou 2008; Zwaan et al. 2002; Louwerse et al. 2006); on the other hand, there are also solid arguments claiming for a primarily linguistic component in comprehension (Louwerse and Jeuniaux 2008). It is not our aim to contribute to this debate, but rather to unite both approaches in a single study which will account for the dual nature of specialized concepts, which seems to be both embodied and linguistic. Previous work in the field has also tried to put both views together (Barsalou et al. 2008; Louwerse and Connell 2011; Lynott and Connell 2010). Embodiment is an approach to cognition, in general, and language comprehension, in particular, according to which
123
Cogn Process
‘‘meaning construction heavily relies on perceptually simulating the information that is presented to the comprehender’’ (Louwerse and Jeuniaux 2010: 96). In other words, comprehension follows perception, for it is the speaker’s bodily experience and environment that trigger the linguistic expressions that carry conceptual meaning to language users. Some of the main assumptions of linguistic relativity and linguistic determinism still underlie some approaches to conceptual processing from a mainly linguistic and symbolic viewpoint. Nevertheless, it should be admitted that words, in so far as ‘‘symbols, can, but do not always have to, be grounded. [In fact,] language is structured in such a way that many relationships that can also be found in the embodied world are structured in language. Language thereby provides a shortcut to the embodied relations in the world’’ (Louwerse and Jeuniaux 2008: 313). Here, we explore the embodied nature of medical concepts through the lexicalization of image schemas in definitions, contexts and terms, and focus on their role in specialized knowledge visualization (SKV). At a cognitive level, image schemas are among the mental representations which, to a great extent, evidence that conceptualization is ‘‘structured in part by the nature of the bodies we have’’ (Evans and Green 2006: 46). In line with Mandler (2004), we think that the basic meaning conveyed by image schemas (such as CONTAIMENT,1 SOURCE-PATHGOAL) can be extended metaphorically to more complex concepts through perceptual meaning analysis and scaffolding. We assume that image schemas underlie certain medical concepts, like PAIN, and that they trigger the production and comprehension of medical language. Our view is that the human conceptual system is grounded in sensorimotor simulations, so that when people hear the word ‘‘dog’’ they more or less vividly bring to memory the animal’s velvety fur, its loud barking and its characteristic smell, etc. Similarly, when people hear painrelated words, such as lumbalgia, neck pain, migraine, headache and sore throat., they will simulate sensorimotor experiences such as the quality of pain (needling, drilling, sharp), its location (chest, back, knee), or its intensity (light, mild, moderate, severe, excruciating); its consequences (manageable, unavoidable, debilitating); and its emotional implications (unbearable, tolerable, indescribable). Our basic hypothesis is that medical pictures based on image schemas lead to better understanding of medical concepts than those which do not depict, either explicitly or implicitly, embodied image-schematic experiences or conceptualizations. In this sense, finding out which images are most representative of medical concepts is paramount 1
In this article capital letters have been used for IMAGE SCHEMAS, small capital letters for CONCEPTS and italics for terms.
123
to provide a consistent selection method for their inclusion in terminological databases. In this article, we study (1) a multimodal corpus of medical texts to examine how image schemas lexicalize in the language of medicine to represent specialized concepts and (2) medical pictures to explore the depiction of imageschematic concepts, in order to account for the verbal and visual representation of embodied concepts. In line with Talmy’s (1983, 2003) research, we aim to explain how language lexicalizes conceptual structure, which in turn reflects embodied experience. This article is divided into six sections. Section ‘‘Embodiment and specialized language: the Frame-based Terminology approach’’ explains the recent cognitive theories of terminology and specialized language from an embodied perspective, paying especial attention to framebased terminology (FbT). Section ‘‘The role of image schemas in specialized knowledge visualization’’ discusses the relationship between image schemas and SKV as it is understood in our study. Section ‘‘Image-schematic representations for pain’’ deals with the way the concept PAIN is linguistically encoded and how the underlying image schemas are lexicalized into English definitions, contexts and terms and visualized in images. In ‘‘A study on the representativeness of image-schematic pictures: preliminary results’’ section, we briefly present preliminary experimental data that provides insights into the role of image schemas in the depiction of medical concepts on the basis of image representativeness. Finally, ‘‘Concluding remarks’’ section lists the conclusions drawn from our research.
Embodiment and specialized language: the frame-based terminology approach Terminology, the study of specialized language and its representation, has begun to pay special attention to cognitive-oriented linguistic theories to study specialized language, according to Faber (Faber 2012). Both cognitive linguistics and terminology have common interests in areas such as the conceptual reference of terminological units, the ontological structure of scientific and technical domains and specialized knowledge representation (see Fig. 1 below). As a result, some approaches have appeared to describe specialized language on a cognitive basis: sociocognitive terminology (Temmerman 2000); termontography (Temmerman and Kerrremans 2003) and FbT (Faber 2012). One of the most interesting aspects of FbT is its focus on the cognitive processing of specialized terms from a dynamic view, as opposed to other theories which only deal with the communicative aspects of terms regardless how concepts are acquired and processed. Therefore, FbT
Cogn Process
Fig. 1 Specialized knowledge representation for FbT
assumes that cognition is typically grounded in multiple ways: simulations, situated action and even bodily states. Although behavioural and neuroimaging studies have shed light on cognitive processing, embodiment theories of language comprehension and social cognition still have to face several challenges (Zwaan 2009). In fact, embodiment and simulated interaction have been explored in general language understanding, but few experiments have dealt with specialized language. FbT is currently considering embodied cognition to fill in this gap, as shown in Fig. 1, where embodied cognition accounts for the relationship between the real world we perceive through our sensorimotor experiences and the internal representations we use to process and communicate our knowledge about the world. From the point of view of FbT, embodied cognition seems to be the consequence of the scaffolded mind (Williams et al. 2009), in accordance with which specialized concepts are built upon simpler, more concrete nonspecialized concepts. Scaffolding then helps link abstract ideas to physical actions in order to integrate new knowledge. Already existing conceptual structures imbue newer concepts with meaning through a scaffolding process where image schemas are thought to play an important role (Mandler 2004). This is crucial for the main purpose of FbT: the design of terminological databases in agreement with our conceptualizations of the world and how concepts are arranged and processed within the human brain’s cognitive system. For Goldman and De Vignemont (2009) embodiment can be interpreted according to body anatomy, bodily activity, bodily content and bodily format given that (1) parts of the body, (2) people’s actions and other bodyrelated traits (e.g. posture), (3) mental representations with bodily contents and (4) mental representations in various
bodily formats or codes have an important causal role in cognition. As an example of how linguistic structure reflects embodied or environmental structure, Louwerse and Zwaan (2009) showed that humans acquire geographical knowledge by means of static/dynamic pictorial representations, verbal descriptions and also through the spatial information coming directly from experience. They were able to prove that language encodes geographical information that language users in turn may use in their understanding of language and the world. They found evidence in text corpora that cities that are located together are debated together; moreover, the frequency with which city names occur in texts should correspond to their population sizes, that is to say, cities that are populated more are debated more. The reason is that language is organized in such a way that it reflects semantic relations in the physical world. In other words, ‘‘prelinguistic conceptual knowledge (e.g. geographical proximity) used when speakers formulate utterances gets translated in linguistic conceptualizations (collocations)’’ (Louwerse and Zwaan 2009: 70). For FbT, embodiment plays a crucial role in knowledge acquisition, representation and transfer, since they are the foundations of communication and social interaction. All we know seems to come from our daily interaction with our own bodies and the surrounding environment, which model the way we store conceptual information in our brain and process data in language production and comprehension. Despite language being the most used system for the transfer of knowledge, multimodal theories of communication (Kress 2010; O’Halloran and Smith 2011) advocate for perceptual stimuli other than those of linguistic nature. During language comprehension, people simulate (or covertly imitate) the form and the meaning of what they hear, and use those representations to predict what is likely to happen next (Pickering and Garrod 2009: 1178). However, from a multimodal point of view, this assertion could also be applied to image and gesture interpretation during communication. Indeed, gestures seem to emerge from perceptual and motor simulations that underlie embodied language and mental imagery (Hostetter and Alibali 2008: 502). This means that linguistic planning involves simulation of visuospatial events which are the bases of mental imagery and language production. Then, gestures, like language, actually arise from embodied thinking. As pointed out above, our conceptual system is not only embodied but also symbolic. Embodiment has also an influence on how we understand the semantic content depicted in images when trying to communicate meaning through visual representations. In this regard, embodiment exerts an influence over the conceptualization of most medical concepts referring to diseases, signs, symptoms and parts of the body, because they are part of our vital experience, and PAIN is no exception. Pain is a highly
123
Cogn Process
unpleasant physical sensation caused by illness or injury (Oxford Dictionary) or an unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage, as defined by the International Association for the Study of Pain. This definition points to the embodied conceptualization of pain, which has been supported by some studies (Melzack and Casey 1968). In fact, the multidimensional theory of pain holds that the conceptualization of pain relies on three key dimensions: sensory, cognitive and affective. The sensory dimension refers to the actual pain sensation transmitted from the pain centre through pain fibres. Besides, the cognitive and affective dimensions refer to the reactions to painful experiences for the sake of pain control and the reduction of anxiety, distress, fear and aversion. Pollatos et al. (2012) have found evidence for a positive relationship between interoceptive sensitivity and the perception of pain stimuli by demonstrating that interoceptive processes and sensitivity to interoceptive signals of bodily arousal (heart rate variability, skin conductance response, baroreflex sensitivity, etc.) are crucial variables for explaining interindividual differences with respect to the perception of pain and its cognitive–affective evaluation. According to Auvray et al. (2010: 214), pain, conceived as an internal perception of bodily damage, is just like exteroceptive perception (vision, audition), with the only difference being that it is not oriented towards publicly available objects, but rather towards events that are taking place in/to one’s own body.
The role of image schemas in specialized knowledge visualization Within FbT, a new approach has been developed to account for the way pictorial representations need to be included in terminological databases and to shed some light into the so far under-researched issue of SKV. Identifying the organization of concepts belonging to a domain on the basis of the conceptual relations established among them is a major concern for designing a terminological database aimed at representing how we think of the outside world by means of concepts. Thus, the aim is to contribute to the holistic representation of concepts by providing criterionreferenced guidelines which support the inclusion in terminological databases of images consistent with the conceptual information contained in definitions and the linguistic information in contexts, for images are nonverbal designations or ‘‘representations of a concept by means other than a descriptive statement, while revealing characteristics of this concept’’ (ISO 10241-1: 2011). Besides descriptive theories of terminology, standardization accounts have evolved to acknowledge the need for the images in terminographical resources, since multimodal
123
content is seen as paramount in these standards. On the one hand, the ISO standard 704 (2000: 23) merely mentions that ‘‘a graphic representation serves its purpose well if it illustrates the characteristics of a given concept and/or its relations to other concepts’’, and it concretizes that images can be iconic, abstract, statistical diagrams or mixed figures. The relation between those images and scientific and technical texts was previously studied by Prieto Velasco (2008, 2013), Prieto Velasco and Lo´pez Rodrı´guez (2009), Prieto Velasco and Faber (2012). On the other hand, the reviewed ISO standard 704 (2009: 46) offers a more detailed explanation of the role of images in terminological databases, as they are considered to be a kind of ostensive definitions. However, it only includes a subtle reference to the notion of multimodality: Also known as a demonstrative definition, is one that defines by exhibiting non-lexical representations of the concept (such as a drawing, an illustration, a video, a sound clip, a computer animation, etc.) or even by pointing to an object. With the increased availability of multimedia technology, ostensive definitions may use any form of multimedia that allows one to exhibit non-lexical representations of the concept. However, rather than being used on their own, ostensive definitions are best employed as complements to intensional definitions or concept descriptions, since it is not always clear what is being referred to or how far to generalize from the particular object exhibited. Furthermore, it may prove difficult to deduce the superordinate concept from an ostensive definition. ISO 704 (2009: 46). By contrast, Galinski and Picht (1997: 55) state that, according to the advances in knowledge representation techniques, ‘‘it is no longer defensible to maintain […] that graphic and pictorial representations should only be used as devices to supplement definitions’’, since they ‘‘can completely assume the function of both terms and definitions’’. At least this is true in certain domains such as medicine, where the communication of knowledge among experts, and between experts and patients is paramount. In fact, the more recent ISO standard 10241-1 (2011: 27) explicitly mentions the non-verbal representations used to exemplify definitions and, to some extent, acknowledges that ‘‘nonverbal representations should not replace a definition but complement it, except in domains or subjects in which nonverbal representations are conventionally used instead of a definition’’. We generally agree with the previous assertion but, either as complements or as alternatives of definitions, images should be selected in a principled way, just like the ISO 10241-1 (2011) describes for the cases, like in medicine, when several preferred, admitted and deprecated
Cogn Process
terms coexist for the sake of terminological variation (Tercedor Sa´nchez 2011). It is our assertion that image schemas, due to their embodied multimodal nature, can be of much help in the representation of specialized knowledge in terminological databases. It is widely agreed that concepts are chunks of knowledge resulting from categorization and that we use them to represent and communicate experiences. If we assume, as many cognitive linguists do (e.g. Cuenca and Hilferty 1999; Mandler 2004; Evans and Green 2006) that (1) there are certain cognitive structures or conceptual patterns which underlie the way we perceive the world and the resulting concepts and (2) that from a semiotic perspective, terms, images and symbols, etc. are just different verbal or non-verbal modes of concept representation, then we can argue that those structures should be found as conceptual primitives when exploring both the linguistic and non-linguistic signs used to represent specialized concepts. Therefore, the SKV approach we described above accounts for the study of textual corpora in order to identify the image schemas underlying a given concept and look for images which consistently depict the embodied experience evoked by such image schemas with a view to their inclusion in terminological databases. The importance of image schemas in cognitive processing is closely related to ‘‘their pervasiveness in experience: to be communicated, our experience must be construed in terms of basic structure, scales and force dynamics’’ (Croft and Cruse 2004: 69). According to Evans and Green (2006: 176–161), image schemas are relatively abstract conceptual representations that arise directly from our everyday interaction with and observation of the world around us. They can be described according to the following characteristics: •
• • • •
Pre-conceptual origin, because they are abstract concepts grounded on sensory experience as the foundations of the conceptual system; Conceptual specificity, because they can give rise to more specific lexical concepts; Experiential nature, because they derive from interaction with and observation of the world; Predictability, because they are inherently meaningful and have predictable consequences; Multimodality, because they can be derived from any sensory modality (sight, touch, hearing or movement/ balance), despite their apparently pictorial nature.
Some concepts like image schemas2 that seem to be (nearly) linguistically universal, like those arising from 2 For an extensive list of image schemas, see Lakoff (1987) and Johnson (1987). For an in-depth discussion on the cognitive and neurobiological grounding of image schemata, see Rohrer (2009: 165–193).
body parts, actions, desires and experiences, etc., are believed to be the foundations of meaning (Aziz-Zadeh et al. 2008). As a result, image schemas can be regarded as embodied multimodal conceptual primitives for language and thought (Huang et al. 2013), since they arise from bodily experience and are deeply grounded in perception and sensorimotor interaction. They can have a visual, aural, olfactory, gustatory or haptic nature and evoke concepts based on exteroception, interoception and proprioception which can be lexicalized through language. Consequently, ‘‘when an abstract concept is scaffolded onto a foundational concept, these concepts become associated much with the same way semantically related concepts are naturally associated with the mind’’ (Williams et al. 2009: 1257). Terminographical hierarchies are an example of such conceptual scaffolding. Concepts are scaffolded on the basis of their specificity, from the most general to the most specific, in ontology-like structures. The more general concepts can then be explained in terms of more basic, simpler concepts acting the role of primitives, whose meaning is extended metaphorically, as we pointed above. The description of specialized domains, according to FbT, is founded on the events that generally take place in them and can be represented accordingly. In the case of medicine in general, and of pain in particular, terms designating diseases are commonly understood as processes on the basis of a SOURCE-PATH-GOAL image schema. Thus, all pathological processes can be somehow linked to a SOURCE or BEGIN OF PATH initiated by an agent; a PATH, which would involve the process itself; and a GOAL or END OF PATH, which would refer to the patient or result of such a process. For instance, AIDS is caused by the human immunodeficiency virus (agent/pathogen), which enters the body through the blood and other bodily fluids (source/begin of path); HIV infects immune cells (goal/patient) leading to low levels of CD4 and T cells through a number of mechanisms (path/process) including apoptosis of uninfected cells, direct viral killing of infected cells and killing of infected CD4? T cells. The extraction of semantic and syntactic information from medical corpora (bottom-up approach) and from specialized dictionaries and experts in the field (top-down approach) contributes to establish terminographical hierarchies (see Table 1) and examine the image schemas behind the lexicalization of medical concepts in definitions and contexts. Besides, SKV does not conceive the inclusion of images in terminological databases as completely detached from other representations. Images are just one among several possible devices for concept representation which need to be fully integrated into terminological databases in accordance with the information provided by other kinds of
123
Cogn Process Table 1 Conceptual hierarchy of
ADENOID CYSTIC CARCINOMA
adenoid cystic carcinoma malignant submandibular gland tumor malignant salivary gland tumor gland tumor malignant tumor cancer
dictionaries and experts, and in knowledge-rich defining contexts extracted from a medical corpus of texts in the form of concordances. Following the recommendations for consistency in the ISO standards, we will use image schemas as the structures at the core of any concept representation (either linguistic or non-linguistic). They appear to activate a common conceptualization which can be designated by terms, definitions, contexts and images.
disease
Lexicalization of image schemas in intensional definitions
natural process process [SOURCE-PATH-GOAL] [PATH]
definitions, which also stand for the concept and reflect the underlying image schema. These definitional representations include: •
•
•
Intensional definitions: definitions which allow the user to recognize and differentiate the concept from other related concepts, providing conceptual information. Defining knowledge-rich contexts: short excerpts that contain the concept designation and allow the user to deduce its meaning by implication, providing linguistic information. Ostensive definitions: demonstrative definitions that define by exhibiting non-lexical representations of the concept, providing visual information.
Finally, the convergence of intensional definitions, defining knowledge-rich contexts and ostensive definitions is deeply rooted in the notion of intersemiotic translation. Multimodal databases need to be internally coherent in order to convey the same meaning when ‘‘translating’’ images into definitions and vice versa. According to multimodality, the different designations of a concept need to be semantically equivalent, despite the fact that the source and target text are semiotically non-equivalent (Gottlieb 2005). Therefore, SKV defends the intersemiotic translation of images into verbal descriptions (ekphrasis), and the deverbalization of definitions in images, so as to achieve semantic equivalence among different conceptual representations.
Image-schematic representations for pain The methodology we propose for studying the lexicalization of image schemas in the language of medicine combines a top-down and bottom-up focus. In this section, we describe both approaches in order to explain the linguistic encoding of image schemas in definitions from specialized
123
Specialized concepts are embodied and linguistic in nature, and intensional definitions may offer valuable conceptual information which is worth examining to analyse the image schemas resulting from conceptualization. Following a topdown approach, we have selected several definitions from specialized sources for the concepts SCIATICA and LOW BACK PAIN, on the one hand, and INFARCTION and ANGINA, on the other, in order to illustrate the differences when processing similar concepts. The definitions have been chosen in accordance with the ISO 10241-1 (2011: 26), that is to say, definitions including ‘‘the superordinate concept to which the concept belongs, and a second part enumerates the delimiting characteristics which distinguish this concept from its coordinate’’. Table 2 shows the definitions for SCIATICA and LOW BACK PAIN, which are sometimes confused as they frequently cooccur in a condition called lumbosciatica or lumbosciatalgia. As highlighted by the previous definitions, the main delimiting characteristic which differentiates sciatica from low back pain is the fact that the sciatic nerve pain radiates down the back through the leg, whereas low back pain is static and remains affecting the lower back. This suggests the existence of two different image schemas. Embodiment in the concept SCIATICA is expressed through the VERTICALITY image schema [UP-DOWN3], whereas the CONTAINMENT image schema [CONTENT] evokes the embodied conceptualization behind LOW BACK PAIN. This is also the case of ANGINA PECTORIS, which is contained within the thoracic cavity, as shown in Table 3. In terms of interoception and proprioception, the difference between angina pectoris and myocardial infarction is the fact that angina pain is, on most occasions, only felt in the chest. On the contrary, pain during an infarction is commonly radiated from the heart to the upper limbs and the neck, thus evoking a SPACE image schema [CENTREPERIPHERY].
3
The image-schemas referred to in this article come originally from the work by Johnson (1987) and Lakoff (1987).
Cogn Process Table 2 Lexicalization of image schemas in the definition of SCIATICA and LOW BACK PAIN SCIATICA
[VERTICALITY] [UP-DOWN]
Pain radiating from the buttock into the thigh, calf and occasionally the foot. Pain felt down the back and lateral aspect of the thigh, leg and foot is often caused by degeneration or displacement of an intervertebral disc, which encroaches upon and irritates a lower lumbar or an upper sacral spinal nerve root. (Concise Medical Dictionary. 2012. Oxford) Irritation or inflammation of the sciatic nerve characterized by a severe pain radiating down the nerve, from the lower back into the leg. (Kent, M. 2012. The Oxford Dictionary of Sports Science & Medicine. Oxford) Pain that radiates along the path of the sciatic nerve, which branches from your lower back through your hips and buttocks and down each leg. Typically, sciatica affects only one side of your body. (Mayo Clinic) LOW BACK PAIN
[CONTAINMENT] [CONTENT]
Localized pain or discomfort in the lumbosacral region of the back. It is often caused by postural defects when the normal relationship between muscles, bones and other tissues is distorted. Low back pain may also be caused by shortening of the hamstrings following vigorous exercise that puts a strain on the back. (Kent, M. 2012. The Oxford Dictionary of Sports Science & Medicine. Oxford) Pain in the lumbar region, of any cause or description. Severe lumbago, of sudden onset, can be due either to a slipped disc or to a strained muscle or ligament. (Concise Medical Dictionary. 2012. Oxford) Acute and/or chronic discomfort in the lumbar region of the spine caused by a number of factors. (Mosby’s Dictionary of Complementary and Alternative Medicine. 2005. Elsevier)
Lexicalization of image schemas in defining contexts Image schemas are the outcome of conceptualization. From the point of view of embodied cognitive processing, they could be considered as the stuff of thought, and as such we should find them behind all kinds of conceptual representations and defining contexts alike. Defining contexts are not definitions, but they offer relevant linguistic information for the deduction of meaning from strings of words which often appear together in texts or collocate. Thanks to the use corpus analysis tools, like Sketch Engine,4 we are able to look thoroughly into how our conceptualizations of the world and experiences are lexicalized in language by browsing a large number of texts. For this purpose, we have searched through a nearly 12 billion-word corpus (EnTenTen12) containing a wide variety of texts ranging from specialized texts (including medical texts) to general texts. We have checked that pain appears in the corpus 125.4 times per million words, which is a relatively high 4 Sketch Engine (Kilgarriff et al. 2004) is a corpus online application which enables queries within large corpora of texts and look into the semantic prosody of words by exploring their grammatical and collocational patterns.
Table 3 Lexicalization of image schemas in the definition of ANGINA and INFARCTION ANGINA
[CONTAINMENT] [CONTENT]
Abrupt onset of pain or crushing sensation in the chest, often provoked by exercise, and caused by narrowing of the coronary arteries and reduced blood supply to the heart. (Lackie, J. 2012. A Dictionary of Biomedicine. Oxford) Chest pain caused by a reduced blood flow to the heart, when the demand for blood by the heart exceeds the supply as during exercise, stress (such as stress induced by a dental visit), or vascular constriction. (Ireland, R. 2012. A Dictionary of Dentistry. Oxford) Chest pain caused by reduced blood flow to the heart muscle. Angina is a symptom of coronary artery disease. Angina is typically described as squeezing, pressure, heaviness, tightness or pain in your chest. (Mayo Clinic) INFARCTION
[SPACE] [CENTRE-PERIPHERY]
Death of part of the heart muscle due to a blockage of the arteries supplying it (coronary thrombosis). It is characterized by sudden severe chest pain which may radiate to the arms and throat. (Ireland, R. 2012. A Dictionary of Dentistry. Oxford) Death of a segment of heart muscle, which follows interruption of its blood supply. Myocardial infarction is usually confined to the left ventricle. The patient experiences a ‘‘heart attack’’: sudden severe chest pain, which may spread to the arms and throat. (Concise Medical Dictionary. 2012. Oxford) Condition characterized by the formation of a dense wedge-shaped block of dead tissue in the myocardium following an interruption to its blood supply. During a severe heart attack, the patient experiences sudden and very severe chest pains, which may spread to the arms and throat. (Kent, M. 2012. The Oxford Dictionary of Sports Science & Medicine. Oxford)
frequency, considering it is not a specialized corpus. The Word Sketch application allows to analyse the collocational patterns of pain by looking at those words which often accompany this term either as objects, subjects or modifiers, as shown in Fig. 2. If we look in more detail into the verbs which co-occur with pain (relieve, ease, alleviate, subside, radiate, persist) by finding short contexts (called concordances) where pain plays the role of either a subject or an object, then it is possible to examine the most frequent collocations. Those contexts also reveal the image schemas underlying most types of pain (angina, sciatica, abdominal, back, chest, headache), as shown in Table 4, EXISTENCE, COUNTERFORCE, SOURCE, REMOVAL OF RESTRAINT, etc. Contextual information is vital to understand the semantic role and relations among concepts and thus identify the image schemas behind them. As a result, conceptual relations like TAKES_PLACE_IN (for processes) or LOCATED_AT (for entities) evoked by verbs like occur often point at the notion of EXISTENCE. The meaning of verbs is crucial when it comes to explore image schemas. The concordances in Table 4 suggest that verbs activate spatial relations during comprehension, as previous
123
Cogn Process
Fig. 2 Collocational patterns for the term pain
Table 4 Lexicalization of image schemas in defining contexts for pain
The most common reason why angina pain occurs is the reduced flow of blood to
EXISTENCE
treatment for that soreness. If sciatica pain occurs during pregnancy after that correct
EXISTENCE
able to better unwind. Nevertheless, if pain persists for many days then its much better succession, but then it’s gone. The longer pains usually persist for at least a day, up
COUNTERFORCE COUNTERFORCE
is always described as sudden and sharp pain originating in the abdominal cavity. The
SOURCE/BEGIN OF PATH
. It has been demonstrated with headache pain arising from the neck and from the facial
SOURCE/BEGIN OF PATH
recommended for cases in which gallbladder pain is severe enough to have prescription or
SCALE
it just wasn’t firing properly, and the pain was fairly intense , radiating down my
SCALE
the usual cause of back pain. This kind of pain is generally mild and relatively short-lived
SCALE
, may also sometimes be a neuralgia, the pain radiating down the inner side of the thigh hypertension who presents with abdominal pain radiating to the back. He’s going to need dissipates within a couple of months. When the pain radiates from a central location, such
VERTICALITY FRONT-BACK CENTRE-PERIPHERY
underwent surgery to relieve chronic back pain , but the procedure paralyzed him from the
REMOVAL OF RESTRAINT
sleep and fluid diet can easily ease the pain and symptoms of Diverticulitis. If you
REMOVAL OF RESTRAINT
problems, and provides treatments to alleviate pain , lack of mobility, and treat disease conditions REMOVAL OF RESTRAINT
123
fight the disease. The lesions heal, the pain subsides within 3 to 5 weeks and in most
REMOVAL
began the program. Within a few days the pain had receded significantly and within two
REMOVAL
disease. It's generally time-limited. Acute pain generally vanishes once the injuries heals
REMOVAL
Cogn Process Table 5 Lexicalization of image schemas in the proposition of PAIN[Vradiate ? prep] SPACE
CONTAINMENT
PATH
LEFT-RIGHT CENTRE-PERIPHERY CENTRE-PERIPHERY UP-DOWN FRONT-BACK NEAR-FAR NEAR-FAR CONTACT CONTAINER IN-OUT IN-OUT SURFACE SOURCE-GOAL
right-sided neck pain. The patient experienced to cold, with intense anxiety. Coldness It has contraction of the chest with the next time, with only groin pain. When menstrual cramps or gas pains? lower back after your current incident or maybe throat of Herpes B infection at a wound include of neck and radiating to upper limbs or for patients who complain of right-sided pain in the thoracic spine and clearly only a few things into considerable zone. The patient subsequently the buttocks and thighs
research by Richardson et al. (2003) showed. Such spatial elements ‘‘could be part of the metaphoric understanding that underlies much of our language, and is rooted in embodied experiences and cultural influences’’ (ibı´d.: 768). This, in turn, provides evidence for the perceptual-motor character of linguistic representations. For instance, the verb to radiate (with rel. freq. in corpus 7.11) implies literal spatial relationships, and is therefore a good example to illustrate the interaction between linguistic conceptual representations with perceptual–spatial processes. As shown in Table 5, not only does the proposition PAIN[Vmotion ? prep] activate spatial image schemas [LEFT–RIGHT, CENTRE-PERIPHERY, UP-DOWN], but also CONTAINMENT and PATH image schemas. The depiction of image schemas in ostensive definitions Similarly to how language encodes image schemas through the lexicalization of embodied experiences, images, insofar as pictorial representations of concepts, constitute the depiction of such knowledge about the world. They are alleged to be demonstrative definitions which convey in a visual format the meaning verbally transmitted through terms, and may supplement, if not replace, conventional definitions. If image schemas underlie some conceptual designations, like intensional definitions and defining contexts, there is no reason to think otherwise in the case of visual representations of specialized concepts or ostensive definitions. By studying the depiction of the image schemas lexicalized in the definitions and contexts of different types of pain, we would guarantee internal coherence between those data categories and ostensive definitions in terminological databases. Then, the degree of representativeness of images appears to be somehow linked to how explicit image schemas are depicted and how easy or difficult is for the viewer to recognize (either consciously or unconsciously) the image schema which lies behind.
pain radiating pain radiating pains radiating pain radiating pain radiating pain radiating pain radiating pain radiating pain radiating pain radiating pain radiating pain radiating pain radiating
to the right periscapular region from head in all directions, duration. from the heart in all directions down a leg, the almost automatic around to the front and back around the arms and legs. Contact away from the bite wound or blisters around chest wall. • Forward through the back, pains radiating into the chest out into other areas of the disc over the right shoulder. A CT scan from the buttock to the foot
In Fig. 3, image a represents the concept SCIATICA by showing the internal anatomy of the sciatic nerve from the lower back to the foot. The irritated nerve is coloured in red and the painful sensation is depicted by arrows showing the direction to which pain radiates. The pictorial representation of sciatica suggests a VERTICALITY image schema [UP-DOWN]. Image B, however, uses concentric circles to indicate that the intensity of LOW BACK PAIN decreases as it goes away from its source, but it seems not to spread. This is the main difference between sciatica and low back pain, as we mentioned above, at least from a sensory-perceptual perspective. The same happens with the depictions of INFARCTION and ANGINA in images C and D, where colour red has conventionally been used to represent painful areas during a cardiac episode. Image C represents the left arm, left part of the chest, neck, jaw and ear to which infarction pain typically spreads. Such a depiction points to a SPACE image schema [CENTRE-PERIPHERY], which is helpful to distinguish myocardial infarction from angina pectoris. Image D represents the chest cavity within which pain is felt during angina. Lighter red is used to depict occasionally referred discomfort in the upper abdomen, shoulders and arms. This pictorial representation evokes a CONTAINMENT image schema, according to which the CONTENT is pain and the CONTAINER is the chest. For the purpose of SKV, the most suitable images for their inclusion in terminological databases would be those which represent the embodied characteristics of the concept depicted, thus facilitating the convergence of the information visually provided by ostensive definitions and the linguistic information provided by intensional definitions and defining contexts. Lexicalization of image schemas in naming concepts We have thus far illustrated how image schemas and mental images shape the way we refer to concepts. These
123
Cogn Process
Concept Image schema
Visual representations
Image schema
SCIATICA
LOW BACK PAIN
(a) UP-DOWN
(b) NEAR
Source: Unknown author http://azpaincenters.wordpress.com/2010/11/05/sid elined-by-low-back-pain/
Source:http://www.ezgaitbrace.com/medicalbracea pplications.html
INFARCTION
ANGINA
(c) CENTREPERIPHERY
(d) IN-OUT
Source: http://www.doctortipster.com/3134myocardial-infarction-causes-risk-factors-signsand-symtoms.html
Visual representations
Source: J. Heuser [GFDL (http://www.gnu.org/copyleft/fdl.html) or CC-BYSA-3.0 (http://creativecommons.org/licenses/bysa/3.0/)], via Wikimedia Commons.
Fig. 3 Visualization of image schemas in the pictorial representation of
SCIATICA, LOW BACK PAIN, INFARCTION
cognitive tools influence the way we perceive and thus conceptualize knowledge. In terminology, the different ways of classifying a reality are called dimensions and concepts are often said to be multidimensional in that they can be named according to the different foci, a result of situatedness and embodiment typical of how we interact with the world, imbued in mental images and image schemas. In VariMed,5 a lexical database containing the different lexicalizations of medical concepts and their motivation, a module has been inserted to include the different dimensions of lexicalizations, i.e. the different ways of conceptualizing particular medical conditions, their signs and symptoms. VariMed database has a twofold purpose: (a) to help communication specialists to choose the most
appropriate term among the different lexicalizations of a specific medical concept, in a given medical context for a specific audience and (b) to serve as a lexical repository of medical variants that researchers in fields such as lexical semantics or applied linguistics can use with different purposes. Dimensions are related to the different ways of seeing (Croft and Cruse 2004) a reality, underlying that ‘‘perceptually […] the same symbols may lead to different classification of the same object’’ (Przybyszewski 2010: 53). Dimensions are thus relations that are made explicit in lexicalizations. In our project, we have arrived to a closed inventory of relations to allow for the classification of diseases and their signs and symptoms: AFFECTS, AGENT/RESULT, VISUAL_ATTRIBUTE_OF, NON_ VISUAL_ATTRIBUTE_OF, MADE_OF, TEMPORAL_ STAGE, INTENSITY, RESEARCHER/DISCOVERER, LOCALIZATION, HAS_ORIGIN_IN. When it comes to defining perception-based concepts, it becomes critical to underline the concept’s most salient
5
VariMed is a terminological database in the field of medicine intended for both researchers and lay people. Its main purposes are research and the popularization of medicine and medical language. It can be accessed at http://varimed.ugr.es.
123
and
ANGINA
Cogn Process Fig. 4 Visually grasping a concept expressed with different image-schematic lexicalizations in English and Spanish
visual attributes; in contrast with other terminological resources, in VariMed images are being recorded to bring to the user’s attention the perceptual features of medical concepts. Specifically, physical symptoms are often best ‘‘visualized’’ when shown in an illustration or a picture, whereas other concepts can do without a visual or ostensive definition. Figure 4 depicts a corkscrew oesophagus, a condition in which uncoordinated contractions of the oesophagus do not propel food effectively to the stomach. The corkscrew mental image is a more elaborated version of the BLOCKAGE image schema. In other words, such image schema is lexicalized in the term corkscrew through a metaphorical extension which leads to a resemblance metaphor which recalls the form of a corkscrew as a VISUAL_ATTRIBUTE_OF. Furthermore, given the importance of situating concepts in a particular experiential frame, defining contexts are being incorporated to illustrate the different elements of the concept’s intension and their relevance in various communication frames. In the following screen, the dimension LOCALIZATION is chosen as the relevant dimension for the term ‘‘abdominal pain’’. Such dimension represents the CONTAINMENT image schema as defined in 4.1, where pain is the CONTENT and abdomen the CONTAINER. Figure 5 shows the series of dimensions which were identified after examining a large corpus of texts. These dimensions are evoked by different term variants on the basis of the different foci highlighted when naming concepts like abdominal pain.
FbT premises posit that terminographical knowledge bases should adjust knowledge activation to the different user and usage needs. As a consequence, it is crucial to acknowledge the conceptual facets which are reflected in definitions (intensional, contexts and ostensive) and the designation of concepts, in order to serve the purpose of user with diverse access profiles.
A study on the representativeness of image-schematic pictures: preliminary results Experimental research is being conducted within the VariMed project to find evidence about the representational value of image-schematic pictures in terminological databases. The scope of the study is to check whether images containing a depiction of the image schemas behind some medical concepts are considered the best pictorial representations by potential users of the VariMed database. It is also our aim to prove whether the degree of specialization of terms has an influence on the type of image deemed as the most representative; in other words, whether popular terms used by lay people to talk about the human body, health and diseases, and specialized terms used by healthcare providers trigger different types of images. For that purpose, we intend to carry out four experiments, the first of which is currently being implemented. We have surveyed English-speaking lay participants and asked them to choose the most representative image of a set of four. They were prompted with a short excerpt
123
Cogn Process Fig. 5 Inventory of dimensions made explicit in the lexicalization of concepts
containing either a highly specialized term or a more explanatory phrase for medical concepts. In the following experiments, we plan to replicate the questionnaire and survey English-speaking healthcare providers. In the second stage, we will survey Spanishspeaking people with no prior medical knowledge and Spanish healthcare providers. This way we will examine the universal nature of image schemas or the potential cultural and linguistic binds (or biases) underlying conceptualization. We hereby present some preliminary results of the first experiment in which we surveyed English-speaking lay people about the most representative images depicting a series of register-based term variants (otalgia-earache, mastectomy-surgical removal of breasts, cephalalgiaheadache, etc.). Given chance performance of 0.25, preliminary results show a statistically significant preference (38.23 %) for image-schematic pictorial representations to the detriment of arbitrarily selected images, which had been chosen at random provided they bore some resemblance with the concept under scrutiny. However, no relevant difference was found in terms of representativeness between imageschematic representations of concepts designated by specialized formal terms (37.20 %), and image-schematic representations of concepts designated by informal popular terms (39.26 %). These figures point to the fact that the depiction of image schemas in medical images contributes
123
to the appropriate pictorial representation of medical concepts, whereas register does not seem to influence to any extent the representativeness of images. Nevertheless, further research is still needed to confirm these findings and be able to compare them with those obtained from healthcare providers in both Spanish and English-speaking contexts.
Concluding remarks We set out to describe how image schemas influence lexicalization of concepts and their definitions, either intensional, ostensive or in the form of contexts. Illustrations, as conceived by FbT, and its implementation in the VariMed database have been used, after a principled selection based on the premises of the SKV approach. Image-schematic illustrations, congruent with the image schemas underlying linguistic representations, seem to be most representative of the concept depicted to the detriment of those images which represent, neither explicitly nor implicitly, the conceptualization of medical knowledge. However, more systematic studies on image representativeness are still needed. Further research will include testing the familiarity of lexicalizations with a Likert-type scale and incorporating corpus-based contexts in all terminological data in order to enhance knowledge transfer and acquisition by user of terminological resources.
Cogn Process Acknowledgments This work was carried out while Dr Prieto Velasco was a visiting researcher at the University of Manchester with the support of a research Grant by the Spanish Ministry of Education (CAS12/00005) in the framework of the R&D project VARIMED (FFI2011-23120). I would like to express my gratitude to Dr Lynott and the colleagues of the Decision and Cognitive Sciences Research Centre and the Embodied Cognition Laboratory at the University of Manchester for their valuable and constructive suggestions to this research work.
References Auvray M, Myin E, Spence C (2010) The sensory-discriminative and affective-motivational aspects of pain. Touch Temp Pleas Pain Special Issue Neurosci Biobehav Rev 34:214–223 Aziz-Zadeh L, Casasanto D, Feldman J, Saxe R, Talmy L (2008) Discovering the conceptual primitives. In: Love BC, McRae K, Sloutsky VM (eds.) Proceedings of the 30th annual conference of the Cognitive Science Society 27–28. Cognitive Science Society, Austin, TX Barsalou LW (2003) Situated simulation in the human conceptual system. Lang Cogn Process 18:513–562 Barsalou LW (2008) Grounded cognition. Annu Rev Psychol 59:617–645 Barsalou LW, Santos A, Simmons WK, Wilson CD (2008) Language and simulation in conceptual processing. In: de Vega M, Glenberg A, Graesser AC (eds) Symbols, embodiment, and meaning. Oxford University Press, Oxford, pp 309–326 Connell L, Lynott D, Dreyer F (2012) A functional role for modalityspecific perceptual systems in conceptual representations. PLoS One 7:3 Croft W, Cruse A (2004) Cognitive linguistics. Cambridge University Press, Cambridge Cuenca MJ, Hilferty J (1999) Introduccio´n a la lingu¨´ıstica cognitiva. Ariel, Barcelona Evans V, Green M (2006) Cognitive linguistics: an introduction. Edinburgh University Press, Edinburgh Faber P (ed) (2012) A cognitive linguistics view of terminology and specialized language. De Gruyter Mouton, New York Galinski C, Picht H (1997) Graphic and other semiotic forms of knowledge representation in terminology management. In: Wright SE, Budin G (eds) Handbook of terminology management, vol 1. John Benjamins, Amsterdam, pp 42–61 Goldman A, De Vignemont F (2009) Is social cognition embodied? Trends Cogn Sci 13(4):154–159 Gottlieb H (2005) Multidimensional translation: semantics turned semiotics. In: Gerzymisch-Arbogast H, Nauert S Proceedings EU high level scientific conferences, Marie Curie Euroconferences MuTra: challenges of multidimensional translation. Saarbru¨cken 2–6 May 2005 Saarland Museum Modern Gallery, Saarbru¨cken Hostetter AB, Alibali M (2008) Visible embodiment: gestures as simulated action. Psychon B Rev 15(3):495–514 Huang X, Huang H, Liao B, Xu C (2013) An ontology-based approach to metaphor cognitive computation. Mind Mach 23:105–121 ISO 10241-1 (2011) Terminological entries in standards. Part 1: general requirements and examples of presentation. International Organization for Standardization, Geneve ISO 704 (2000) Terminology work: principles and methods. International Organization for Standardization, Geneve ISO 704 (2009) Terminology work: principles and methods. International Organization for Standardization, Geneve
Johnson M (1987) The body in the mind: the bodily basis of meaning, imagination, and reason. University of Chicago Press, Chicago Kilgarriff A, Rychly P, Smrz P, Tugwell D (2004) The Sketch Engine. Proc EURALEX 2004, Lorient, France; 105–116. http://www. sketchengine.co.uk Kress G (2010) Multimodality: a social semiotic approach to contemporary communication. Routledge, London Lakoff G (1987) Women, fire, and dangerous things: what categories reveal about the mind. University of Chicago Press, Chicago Louwerse M, Connell L (2011) A taste of words: linguistic context and perceptual simulation predict the modality of words. Cognitive Sci 35(2):381–398 Louwerse M, Jeuniaux P (2008) Language comprehension is both embodied and symbolic. In: de Vega M, Glenberg A, Graesser AC (eds) Symbols, embodiment, and meaning. Oxford University Press, Oxford, pp 309–326 Louwerse M, Jeuniaux P (2010) The linguistic and embodied nature of conceptual processing. Cognition 114:96–104 Louwerse M, Zwaan R (2009) Language encodes geographical information. Cognitive Sci 33:51–73 Louwerse M, Cai Z, Hu X, Ventura M, Jeuniaux P (2006) Cognitively inspired natural language based knowledge representations: further explorations of Latent semantic analysis. Int J Artif Intell Tools 15(1021):1–20 Lynott D, Connell L (2010) Embodied conceptual combination. Front Psychol 1(212):1–14 Mandler JM (2004) The foundations of mind: origins of conceptual thought. Oxford University Press, San Diego Melzack R, Casey KL (1968) Sensory, motivational and central control determinants of chronic pain: a new conceptual model. In: Kenshalo DR (ed) The skin senses: Proceedings of the first international symposium on the skin senses. Florida State University in Tallahassee, Florida O’Halloran KL, Smith BA (eds) (2011) Multimodal studies: exploring issues and domains. Routledge, New York Pickering MJ, Garrod S (2009) Prediction and embodiment in dialogue. Eur J Soc Psychol 39:1162–1168 Pollatos O, Fu¨sto¨s J, Critchley HD (2012) On the generalised embodiment of pain: how interoceptive sensitivity modulates cutaneous pain perception. Pain 153:1680–1686 Prieto Velasco JA (2008) Informacio´n gra´fica y grados de especialidad en el discurso cientı´fico-te´cnico: un estudio de corpus. Dissertation, University of Granada Prieto Velasco JA (2013) A corpus-based approach to the multimodal analysis of specialized knowledge. Lang Resour Eval 47(2): 399–423 Prieto Velasco JA, Faber P (2012) Graphical information. In: Faber P (ed) A cognitive linguistics view of terminology and specialized language. De Gruyter Mouton, Boston, pp 225–248 Prieto Velasco JA, Lo´pez Rodrı´guez CI (2009) Managing graphic information in terminological knowledge bases. Terminology 15(2):179–213 Przybyszewski AW (2010) Logical rules of visual brain: from anatomy through neurophysiology to cognition. Cogn Syst Res 11(1):53–66 Richardson DC, Spivey MJ, Barsalou LW, MCRae K (2003) Spatial representations activated during real time comprehension of verbs. Cogn Sci 27:767–780 Rohrer T (2009) Image schemata in the brain. In: Hampe B (ed) From perception to meaning: image schemas in cognitive linguistics. Mouton De Gruyer, NewYork Talmy L (1983) How language structures space. In: Herbert L, Pick Jr, Acredolo L (eds) Spatial orientation: theory, research, and application, Plenum Pres, New York Talmy L (2003) The representation of spatial structure in spoken and signed language. In: Emmor K (ed) Perspectives on classifier
123
Cogn Process constructions in Sign language. Lawrence Erlbaum, Mahwah, NJ, pp 69–195 Temmerman R (2000) Towards new ways in terminology description. The sociocognitive approach. Amsterdam, John Benjamins Temmerman R, Kerrremans K (2003) Termontography: ontology building and the sociocognitive approach to terminology description. In: Hajicˇova´, E, Koteˇsˇovcova´ A, Mı´rovsky´ J. (eds) Proceedings of CIL17, Matfyzpress, MFF UK (CD-ROM). Prague, Czech Republic Tercedor Sa´nchez M (2011) The cognitive dynamics of terminological variation. Terminology 17(2):181–197
123
Tercedor Sa´nchez M, Uren˜a Go´mez-Moreno JM, Prieto Velasco JA (2012) Metaphoric and metonymic processes in terminology. J Spec Transl 18:187–205 Williams LW, Huang JY, Bargh JA (2009) The scaffolded mind: higher mental processes are grounded in early experience of the physical world. Eur J Soc Psychol 39:1257–1267 Zwaan RA (2009) Mental simulation in language comprehension and social cognition. Eur J Soc Psychol 39:1142–1150 Zwaan RA, Stanfield RA, Yaxley RH (2002) Language comprehenders mentally represent the shape of objects. Psychol Sci 13:168–171
RESEARCH REPORT
The embodied nature of medical concepts: image schemas and language for PAIN Juan Antonio Prieto Velasco Maribel Tercedor Sa´nchez
•
Received: 15 July 2013 / Accepted: 12 December 2013 Marta Olivetti Belardinelli and Springer-Verlag Berlin Heidelberg 2014
Abstract Cognitive linguistics assumes that knowledge is both embodied and situated as far as it is acquired through our bodily interaction with the world in a specific environment (e.g. Barsalou in Lang Cogn Process 18:513–562, 2003; Connell et al. in PLoS One 7:3, 2012). Therefore, embodiment provides an explanation to the mental representation and linguistic expression of concepts. Among the first, we find multimodal conceptual structures, like image schemas, which are schematic representations of embodied experiences resulting from our conceptualization of the surrounding environment (Tercedor Sa´nchez et al. in J Spec Transl 18:187–205, 2012). Furthermore, the way we interact with the environment and its objects is dynamic and configures how we refer to concepts both by means of images and lexicalizations. In this article, we investigate how image schemas underlie verbal and visual representations. They both evoke concepts based on exteroception, interoception and proprioception which can be lexicalized through language. More specifically, we study (1) a multimodal corpus of medical texts to examine how image schemas lexicalize in the language of medicine to represent specialized concepts and (2) medical pictures to explore the depiction of image-schematic concepts, in order to account This article is part of the Special Section on ‘‘Embodied Social Cognition,’’ guest-edited by Fernando Marmolejo Ramos and Amedeo Dangiulli. J. A. Prieto Velasco (&) Department of Philology and Translation, University Pablo de Olavide, Ctra. de Utrera km. 1, 41013 Seville, Spain e-mail: [email protected] M. Tercedor Sa´nchez Department of Translation and Interpreting, University of Granada, C/Buensuceso, 11, 18002 Granada, Spain e-mail: [email protected]
for the verbal and visual representation of embodied concepts. We explore the concept PAIN, a sensory and emotional experience associated with actual or potential tissue damage, using corpus analysis tools (Sketch Engine) to extract information about the lexicalization of underlying image schemas in definitions and defining contexts. Then, we use the image schemas behind medical concepts to consistently select images which depict our experience of pain and the way we understand it. Finally, such lexicalizations and visualizations will help us assess how we refer to PAIN both verbally and visually. Keywords Image schemas Embodiment Knowledge visualization Medical concepts
Introduction Research on cognition and language understanding has paid close attention to the discussion about whether conceptual processing is primarily embodied or linguistic. On the one hand, there is experimental evidence for the existence of a major embodied component (Barsalou 2008; Zwaan et al. 2002; Louwerse et al. 2006); on the other hand, there are also solid arguments claiming for a primarily linguistic component in comprehension (Louwerse and Jeuniaux 2008). It is not our aim to contribute to this debate, but rather to unite both approaches in a single study which will account for the dual nature of specialized concepts, which seems to be both embodied and linguistic. Previous work in the field has also tried to put both views together (Barsalou et al. 2008; Louwerse and Connell 2011; Lynott and Connell 2010). Embodiment is an approach to cognition, in general, and language comprehension, in particular, according to which
123
Cogn Process
‘‘meaning construction heavily relies on perceptually simulating the information that is presented to the comprehender’’ (Louwerse and Jeuniaux 2010: 96). In other words, comprehension follows perception, for it is the speaker’s bodily experience and environment that trigger the linguistic expressions that carry conceptual meaning to language users. Some of the main assumptions of linguistic relativity and linguistic determinism still underlie some approaches to conceptual processing from a mainly linguistic and symbolic viewpoint. Nevertheless, it should be admitted that words, in so far as ‘‘symbols, can, but do not always have to, be grounded. [In fact,] language is structured in such a way that many relationships that can also be found in the embodied world are structured in language. Language thereby provides a shortcut to the embodied relations in the world’’ (Louwerse and Jeuniaux 2008: 313). Here, we explore the embodied nature of medical concepts through the lexicalization of image schemas in definitions, contexts and terms, and focus on their role in specialized knowledge visualization (SKV). At a cognitive level, image schemas are among the mental representations which, to a great extent, evidence that conceptualization is ‘‘structured in part by the nature of the bodies we have’’ (Evans and Green 2006: 46). In line with Mandler (2004), we think that the basic meaning conveyed by image schemas (such as CONTAIMENT,1 SOURCE-PATHGOAL) can be extended metaphorically to more complex concepts through perceptual meaning analysis and scaffolding. We assume that image schemas underlie certain medical concepts, like PAIN, and that they trigger the production and comprehension of medical language. Our view is that the human conceptual system is grounded in sensorimotor simulations, so that when people hear the word ‘‘dog’’ they more or less vividly bring to memory the animal’s velvety fur, its loud barking and its characteristic smell, etc. Similarly, when people hear painrelated words, such as lumbalgia, neck pain, migraine, headache and sore throat., they will simulate sensorimotor experiences such as the quality of pain (needling, drilling, sharp), its location (chest, back, knee), or its intensity (light, mild, moderate, severe, excruciating); its consequences (manageable, unavoidable, debilitating); and its emotional implications (unbearable, tolerable, indescribable). Our basic hypothesis is that medical pictures based on image schemas lead to better understanding of medical concepts than those which do not depict, either explicitly or implicitly, embodied image-schematic experiences or conceptualizations. In this sense, finding out which images are most representative of medical concepts is paramount 1
In this article capital letters have been used for IMAGE SCHEMAS, small capital letters for CONCEPTS and italics for terms.
123
to provide a consistent selection method for their inclusion in terminological databases. In this article, we study (1) a multimodal corpus of medical texts to examine how image schemas lexicalize in the language of medicine to represent specialized concepts and (2) medical pictures to explore the depiction of imageschematic concepts, in order to account for the verbal and visual representation of embodied concepts. In line with Talmy’s (1983, 2003) research, we aim to explain how language lexicalizes conceptual structure, which in turn reflects embodied experience. This article is divided into six sections. Section ‘‘Embodiment and specialized language: the Frame-based Terminology approach’’ explains the recent cognitive theories of terminology and specialized language from an embodied perspective, paying especial attention to framebased terminology (FbT). Section ‘‘The role of image schemas in specialized knowledge visualization’’ discusses the relationship between image schemas and SKV as it is understood in our study. Section ‘‘Image-schematic representations for pain’’ deals with the way the concept PAIN is linguistically encoded and how the underlying image schemas are lexicalized into English definitions, contexts and terms and visualized in images. In ‘‘A study on the representativeness of image-schematic pictures: preliminary results’’ section, we briefly present preliminary experimental data that provides insights into the role of image schemas in the depiction of medical concepts on the basis of image representativeness. Finally, ‘‘Concluding remarks’’ section lists the conclusions drawn from our research.
Embodiment and specialized language: the frame-based terminology approach Terminology, the study of specialized language and its representation, has begun to pay special attention to cognitive-oriented linguistic theories to study specialized language, according to Faber (Faber 2012). Both cognitive linguistics and terminology have common interests in areas such as the conceptual reference of terminological units, the ontological structure of scientific and technical domains and specialized knowledge representation (see Fig. 1 below). As a result, some approaches have appeared to describe specialized language on a cognitive basis: sociocognitive terminology (Temmerman 2000); termontography (Temmerman and Kerrremans 2003) and FbT (Faber 2012). One of the most interesting aspects of FbT is its focus on the cognitive processing of specialized terms from a dynamic view, as opposed to other theories which only deal with the communicative aspects of terms regardless how concepts are acquired and processed. Therefore, FbT
Cogn Process
Fig. 1 Specialized knowledge representation for FbT
assumes that cognition is typically grounded in multiple ways: simulations, situated action and even bodily states. Although behavioural and neuroimaging studies have shed light on cognitive processing, embodiment theories of language comprehension and social cognition still have to face several challenges (Zwaan 2009). In fact, embodiment and simulated interaction have been explored in general language understanding, but few experiments have dealt with specialized language. FbT is currently considering embodied cognition to fill in this gap, as shown in Fig. 1, where embodied cognition accounts for the relationship between the real world we perceive through our sensorimotor experiences and the internal representations we use to process and communicate our knowledge about the world. From the point of view of FbT, embodied cognition seems to be the consequence of the scaffolded mind (Williams et al. 2009), in accordance with which specialized concepts are built upon simpler, more concrete nonspecialized concepts. Scaffolding then helps link abstract ideas to physical actions in order to integrate new knowledge. Already existing conceptual structures imbue newer concepts with meaning through a scaffolding process where image schemas are thought to play an important role (Mandler 2004). This is crucial for the main purpose of FbT: the design of terminological databases in agreement with our conceptualizations of the world and how concepts are arranged and processed within the human brain’s cognitive system. For Goldman and De Vignemont (2009) embodiment can be interpreted according to body anatomy, bodily activity, bodily content and bodily format given that (1) parts of the body, (2) people’s actions and other bodyrelated traits (e.g. posture), (3) mental representations with bodily contents and (4) mental representations in various
bodily formats or codes have an important causal role in cognition. As an example of how linguistic structure reflects embodied or environmental structure, Louwerse and Zwaan (2009) showed that humans acquire geographical knowledge by means of static/dynamic pictorial representations, verbal descriptions and also through the spatial information coming directly from experience. They were able to prove that language encodes geographical information that language users in turn may use in their understanding of language and the world. They found evidence in text corpora that cities that are located together are debated together; moreover, the frequency with which city names occur in texts should correspond to their population sizes, that is to say, cities that are populated more are debated more. The reason is that language is organized in such a way that it reflects semantic relations in the physical world. In other words, ‘‘prelinguistic conceptual knowledge (e.g. geographical proximity) used when speakers formulate utterances gets translated in linguistic conceptualizations (collocations)’’ (Louwerse and Zwaan 2009: 70). For FbT, embodiment plays a crucial role in knowledge acquisition, representation and transfer, since they are the foundations of communication and social interaction. All we know seems to come from our daily interaction with our own bodies and the surrounding environment, which model the way we store conceptual information in our brain and process data in language production and comprehension. Despite language being the most used system for the transfer of knowledge, multimodal theories of communication (Kress 2010; O’Halloran and Smith 2011) advocate for perceptual stimuli other than those of linguistic nature. During language comprehension, people simulate (or covertly imitate) the form and the meaning of what they hear, and use those representations to predict what is likely to happen next (Pickering and Garrod 2009: 1178). However, from a multimodal point of view, this assertion could also be applied to image and gesture interpretation during communication. Indeed, gestures seem to emerge from perceptual and motor simulations that underlie embodied language and mental imagery (Hostetter and Alibali 2008: 502). This means that linguistic planning involves simulation of visuospatial events which are the bases of mental imagery and language production. Then, gestures, like language, actually arise from embodied thinking. As pointed out above, our conceptual system is not only embodied but also symbolic. Embodiment has also an influence on how we understand the semantic content depicted in images when trying to communicate meaning through visual representations. In this regard, embodiment exerts an influence over the conceptualization of most medical concepts referring to diseases, signs, symptoms and parts of the body, because they are part of our vital experience, and PAIN is no exception. Pain is a highly
123
Cogn Process
unpleasant physical sensation caused by illness or injury (Oxford Dictionary) or an unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage, as defined by the International Association for the Study of Pain. This definition points to the embodied conceptualization of pain, which has been supported by some studies (Melzack and Casey 1968). In fact, the multidimensional theory of pain holds that the conceptualization of pain relies on three key dimensions: sensory, cognitive and affective. The sensory dimension refers to the actual pain sensation transmitted from the pain centre through pain fibres. Besides, the cognitive and affective dimensions refer to the reactions to painful experiences for the sake of pain control and the reduction of anxiety, distress, fear and aversion. Pollatos et al. (2012) have found evidence for a positive relationship between interoceptive sensitivity and the perception of pain stimuli by demonstrating that interoceptive processes and sensitivity to interoceptive signals of bodily arousal (heart rate variability, skin conductance response, baroreflex sensitivity, etc.) are crucial variables for explaining interindividual differences with respect to the perception of pain and its cognitive–affective evaluation. According to Auvray et al. (2010: 214), pain, conceived as an internal perception of bodily damage, is just like exteroceptive perception (vision, audition), with the only difference being that it is not oriented towards publicly available objects, but rather towards events that are taking place in/to one’s own body.
The role of image schemas in specialized knowledge visualization Within FbT, a new approach has been developed to account for the way pictorial representations need to be included in terminological databases and to shed some light into the so far under-researched issue of SKV. Identifying the organization of concepts belonging to a domain on the basis of the conceptual relations established among them is a major concern for designing a terminological database aimed at representing how we think of the outside world by means of concepts. Thus, the aim is to contribute to the holistic representation of concepts by providing criterionreferenced guidelines which support the inclusion in terminological databases of images consistent with the conceptual information contained in definitions and the linguistic information in contexts, for images are nonverbal designations or ‘‘representations of a concept by means other than a descriptive statement, while revealing characteristics of this concept’’ (ISO 10241-1: 2011). Besides descriptive theories of terminology, standardization accounts have evolved to acknowledge the need for the images in terminographical resources, since multimodal
123
content is seen as paramount in these standards. On the one hand, the ISO standard 704 (2000: 23) merely mentions that ‘‘a graphic representation serves its purpose well if it illustrates the characteristics of a given concept and/or its relations to other concepts’’, and it concretizes that images can be iconic, abstract, statistical diagrams or mixed figures. The relation between those images and scientific and technical texts was previously studied by Prieto Velasco (2008, 2013), Prieto Velasco and Lo´pez Rodrı´guez (2009), Prieto Velasco and Faber (2012). On the other hand, the reviewed ISO standard 704 (2009: 46) offers a more detailed explanation of the role of images in terminological databases, as they are considered to be a kind of ostensive definitions. However, it only includes a subtle reference to the notion of multimodality: Also known as a demonstrative definition, is one that defines by exhibiting non-lexical representations of the concept (such as a drawing, an illustration, a video, a sound clip, a computer animation, etc.) or even by pointing to an object. With the increased availability of multimedia technology, ostensive definitions may use any form of multimedia that allows one to exhibit non-lexical representations of the concept. However, rather than being used on their own, ostensive definitions are best employed as complements to intensional definitions or concept descriptions, since it is not always clear what is being referred to or how far to generalize from the particular object exhibited. Furthermore, it may prove difficult to deduce the superordinate concept from an ostensive definition. ISO 704 (2009: 46). By contrast, Galinski and Picht (1997: 55) state that, according to the advances in knowledge representation techniques, ‘‘it is no longer defensible to maintain […] that graphic and pictorial representations should only be used as devices to supplement definitions’’, since they ‘‘can completely assume the function of both terms and definitions’’. At least this is true in certain domains such as medicine, where the communication of knowledge among experts, and between experts and patients is paramount. In fact, the more recent ISO standard 10241-1 (2011: 27) explicitly mentions the non-verbal representations used to exemplify definitions and, to some extent, acknowledges that ‘‘nonverbal representations should not replace a definition but complement it, except in domains or subjects in which nonverbal representations are conventionally used instead of a definition’’. We generally agree with the previous assertion but, either as complements or as alternatives of definitions, images should be selected in a principled way, just like the ISO 10241-1 (2011) describes for the cases, like in medicine, when several preferred, admitted and deprecated
Cogn Process
terms coexist for the sake of terminological variation (Tercedor Sa´nchez 2011). It is our assertion that image schemas, due to their embodied multimodal nature, can be of much help in the representation of specialized knowledge in terminological databases. It is widely agreed that concepts are chunks of knowledge resulting from categorization and that we use them to represent and communicate experiences. If we assume, as many cognitive linguists do (e.g. Cuenca and Hilferty 1999; Mandler 2004; Evans and Green 2006) that (1) there are certain cognitive structures or conceptual patterns which underlie the way we perceive the world and the resulting concepts and (2) that from a semiotic perspective, terms, images and symbols, etc. are just different verbal or non-verbal modes of concept representation, then we can argue that those structures should be found as conceptual primitives when exploring both the linguistic and non-linguistic signs used to represent specialized concepts. Therefore, the SKV approach we described above accounts for the study of textual corpora in order to identify the image schemas underlying a given concept and look for images which consistently depict the embodied experience evoked by such image schemas with a view to their inclusion in terminological databases. The importance of image schemas in cognitive processing is closely related to ‘‘their pervasiveness in experience: to be communicated, our experience must be construed in terms of basic structure, scales and force dynamics’’ (Croft and Cruse 2004: 69). According to Evans and Green (2006: 176–161), image schemas are relatively abstract conceptual representations that arise directly from our everyday interaction with and observation of the world around us. They can be described according to the following characteristics: •
• • • •
Pre-conceptual origin, because they are abstract concepts grounded on sensory experience as the foundations of the conceptual system; Conceptual specificity, because they can give rise to more specific lexical concepts; Experiential nature, because they derive from interaction with and observation of the world; Predictability, because they are inherently meaningful and have predictable consequences; Multimodality, because they can be derived from any sensory modality (sight, touch, hearing or movement/ balance), despite their apparently pictorial nature.
Some concepts like image schemas2 that seem to be (nearly) linguistically universal, like those arising from 2 For an extensive list of image schemas, see Lakoff (1987) and Johnson (1987). For an in-depth discussion on the cognitive and neurobiological grounding of image schemata, see Rohrer (2009: 165–193).
body parts, actions, desires and experiences, etc., are believed to be the foundations of meaning (Aziz-Zadeh et al. 2008). As a result, image schemas can be regarded as embodied multimodal conceptual primitives for language and thought (Huang et al. 2013), since they arise from bodily experience and are deeply grounded in perception and sensorimotor interaction. They can have a visual, aural, olfactory, gustatory or haptic nature and evoke concepts based on exteroception, interoception and proprioception which can be lexicalized through language. Consequently, ‘‘when an abstract concept is scaffolded onto a foundational concept, these concepts become associated much with the same way semantically related concepts are naturally associated with the mind’’ (Williams et al. 2009: 1257). Terminographical hierarchies are an example of such conceptual scaffolding. Concepts are scaffolded on the basis of their specificity, from the most general to the most specific, in ontology-like structures. The more general concepts can then be explained in terms of more basic, simpler concepts acting the role of primitives, whose meaning is extended metaphorically, as we pointed above. The description of specialized domains, according to FbT, is founded on the events that generally take place in them and can be represented accordingly. In the case of medicine in general, and of pain in particular, terms designating diseases are commonly understood as processes on the basis of a SOURCE-PATH-GOAL image schema. Thus, all pathological processes can be somehow linked to a SOURCE or BEGIN OF PATH initiated by an agent; a PATH, which would involve the process itself; and a GOAL or END OF PATH, which would refer to the patient or result of such a process. For instance, AIDS is caused by the human immunodeficiency virus (agent/pathogen), which enters the body through the blood and other bodily fluids (source/begin of path); HIV infects immune cells (goal/patient) leading to low levels of CD4 and T cells through a number of mechanisms (path/process) including apoptosis of uninfected cells, direct viral killing of infected cells and killing of infected CD4? T cells. The extraction of semantic and syntactic information from medical corpora (bottom-up approach) and from specialized dictionaries and experts in the field (top-down approach) contributes to establish terminographical hierarchies (see Table 1) and examine the image schemas behind the lexicalization of medical concepts in definitions and contexts. Besides, SKV does not conceive the inclusion of images in terminological databases as completely detached from other representations. Images are just one among several possible devices for concept representation which need to be fully integrated into terminological databases in accordance with the information provided by other kinds of
123
Cogn Process Table 1 Conceptual hierarchy of
ADENOID CYSTIC CARCINOMA
adenoid cystic carcinoma malignant submandibular gland tumor malignant salivary gland tumor gland tumor malignant tumor cancer
dictionaries and experts, and in knowledge-rich defining contexts extracted from a medical corpus of texts in the form of concordances. Following the recommendations for consistency in the ISO standards, we will use image schemas as the structures at the core of any concept representation (either linguistic or non-linguistic). They appear to activate a common conceptualization which can be designated by terms, definitions, contexts and images.
disease
Lexicalization of image schemas in intensional definitions
natural process process [SOURCE-PATH-GOAL] [PATH]
definitions, which also stand for the concept and reflect the underlying image schema. These definitional representations include: •
•
•
Intensional definitions: definitions which allow the user to recognize and differentiate the concept from other related concepts, providing conceptual information. Defining knowledge-rich contexts: short excerpts that contain the concept designation and allow the user to deduce its meaning by implication, providing linguistic information. Ostensive definitions: demonstrative definitions that define by exhibiting non-lexical representations of the concept, providing visual information.
Finally, the convergence of intensional definitions, defining knowledge-rich contexts and ostensive definitions is deeply rooted in the notion of intersemiotic translation. Multimodal databases need to be internally coherent in order to convey the same meaning when ‘‘translating’’ images into definitions and vice versa. According to multimodality, the different designations of a concept need to be semantically equivalent, despite the fact that the source and target text are semiotically non-equivalent (Gottlieb 2005). Therefore, SKV defends the intersemiotic translation of images into verbal descriptions (ekphrasis), and the deverbalization of definitions in images, so as to achieve semantic equivalence among different conceptual representations.
Image-schematic representations for pain The methodology we propose for studying the lexicalization of image schemas in the language of medicine combines a top-down and bottom-up focus. In this section, we describe both approaches in order to explain the linguistic encoding of image schemas in definitions from specialized
123
Specialized concepts are embodied and linguistic in nature, and intensional definitions may offer valuable conceptual information which is worth examining to analyse the image schemas resulting from conceptualization. Following a topdown approach, we have selected several definitions from specialized sources for the concepts SCIATICA and LOW BACK PAIN, on the one hand, and INFARCTION and ANGINA, on the other, in order to illustrate the differences when processing similar concepts. The definitions have been chosen in accordance with the ISO 10241-1 (2011: 26), that is to say, definitions including ‘‘the superordinate concept to which the concept belongs, and a second part enumerates the delimiting characteristics which distinguish this concept from its coordinate’’. Table 2 shows the definitions for SCIATICA and LOW BACK PAIN, which are sometimes confused as they frequently cooccur in a condition called lumbosciatica or lumbosciatalgia. As highlighted by the previous definitions, the main delimiting characteristic which differentiates sciatica from low back pain is the fact that the sciatic nerve pain radiates down the back through the leg, whereas low back pain is static and remains affecting the lower back. This suggests the existence of two different image schemas. Embodiment in the concept SCIATICA is expressed through the VERTICALITY image schema [UP-DOWN3], whereas the CONTAINMENT image schema [CONTENT] evokes the embodied conceptualization behind LOW BACK PAIN. This is also the case of ANGINA PECTORIS, which is contained within the thoracic cavity, as shown in Table 3. In terms of interoception and proprioception, the difference between angina pectoris and myocardial infarction is the fact that angina pain is, on most occasions, only felt in the chest. On the contrary, pain during an infarction is commonly radiated from the heart to the upper limbs and the neck, thus evoking a SPACE image schema [CENTREPERIPHERY].
3
The image-schemas referred to in this article come originally from the work by Johnson (1987) and Lakoff (1987).
Cogn Process Table 2 Lexicalization of image schemas in the definition of SCIATICA and LOW BACK PAIN SCIATICA
[VERTICALITY] [UP-DOWN]
Pain radiating from the buttock into the thigh, calf and occasionally the foot. Pain felt down the back and lateral aspect of the thigh, leg and foot is often caused by degeneration or displacement of an intervertebral disc, which encroaches upon and irritates a lower lumbar or an upper sacral spinal nerve root. (Concise Medical Dictionary. 2012. Oxford) Irritation or inflammation of the sciatic nerve characterized by a severe pain radiating down the nerve, from the lower back into the leg. (Kent, M. 2012. The Oxford Dictionary of Sports Science & Medicine. Oxford) Pain that radiates along the path of the sciatic nerve, which branches from your lower back through your hips and buttocks and down each leg. Typically, sciatica affects only one side of your body. (Mayo Clinic) LOW BACK PAIN
[CONTAINMENT] [CONTENT]
Localized pain or discomfort in the lumbosacral region of the back. It is often caused by postural defects when the normal relationship between muscles, bones and other tissues is distorted. Low back pain may also be caused by shortening of the hamstrings following vigorous exercise that puts a strain on the back. (Kent, M. 2012. The Oxford Dictionary of Sports Science & Medicine. Oxford) Pain in the lumbar region, of any cause or description. Severe lumbago, of sudden onset, can be due either to a slipped disc or to a strained muscle or ligament. (Concise Medical Dictionary. 2012. Oxford) Acute and/or chronic discomfort in the lumbar region of the spine caused by a number of factors. (Mosby’s Dictionary of Complementary and Alternative Medicine. 2005. Elsevier)
Lexicalization of image schemas in defining contexts Image schemas are the outcome of conceptualization. From the point of view of embodied cognitive processing, they could be considered as the stuff of thought, and as such we should find them behind all kinds of conceptual representations and defining contexts alike. Defining contexts are not definitions, but they offer relevant linguistic information for the deduction of meaning from strings of words which often appear together in texts or collocate. Thanks to the use corpus analysis tools, like Sketch Engine,4 we are able to look thoroughly into how our conceptualizations of the world and experiences are lexicalized in language by browsing a large number of texts. For this purpose, we have searched through a nearly 12 billion-word corpus (EnTenTen12) containing a wide variety of texts ranging from specialized texts (including medical texts) to general texts. We have checked that pain appears in the corpus 125.4 times per million words, which is a relatively high 4 Sketch Engine (Kilgarriff et al. 2004) is a corpus online application which enables queries within large corpora of texts and look into the semantic prosody of words by exploring their grammatical and collocational patterns.
Table 3 Lexicalization of image schemas in the definition of ANGINA and INFARCTION ANGINA
[CONTAINMENT] [CONTENT]
Abrupt onset of pain or crushing sensation in the chest, often provoked by exercise, and caused by narrowing of the coronary arteries and reduced blood supply to the heart. (Lackie, J. 2012. A Dictionary of Biomedicine. Oxford) Chest pain caused by a reduced blood flow to the heart, when the demand for blood by the heart exceeds the supply as during exercise, stress (such as stress induced by a dental visit), or vascular constriction. (Ireland, R. 2012. A Dictionary of Dentistry. Oxford) Chest pain caused by reduced blood flow to the heart muscle. Angina is a symptom of coronary artery disease. Angina is typically described as squeezing, pressure, heaviness, tightness or pain in your chest. (Mayo Clinic) INFARCTION
[SPACE] [CENTRE-PERIPHERY]
Death of part of the heart muscle due to a blockage of the arteries supplying it (coronary thrombosis). It is characterized by sudden severe chest pain which may radiate to the arms and throat. (Ireland, R. 2012. A Dictionary of Dentistry. Oxford) Death of a segment of heart muscle, which follows interruption of its blood supply. Myocardial infarction is usually confined to the left ventricle. The patient experiences a ‘‘heart attack’’: sudden severe chest pain, which may spread to the arms and throat. (Concise Medical Dictionary. 2012. Oxford) Condition characterized by the formation of a dense wedge-shaped block of dead tissue in the myocardium following an interruption to its blood supply. During a severe heart attack, the patient experiences sudden and very severe chest pains, which may spread to the arms and throat. (Kent, M. 2012. The Oxford Dictionary of Sports Science & Medicine. Oxford)
frequency, considering it is not a specialized corpus. The Word Sketch application allows to analyse the collocational patterns of pain by looking at those words which often accompany this term either as objects, subjects or modifiers, as shown in Fig. 2. If we look in more detail into the verbs which co-occur with pain (relieve, ease, alleviate, subside, radiate, persist) by finding short contexts (called concordances) where pain plays the role of either a subject or an object, then it is possible to examine the most frequent collocations. Those contexts also reveal the image schemas underlying most types of pain (angina, sciatica, abdominal, back, chest, headache), as shown in Table 4, EXISTENCE, COUNTERFORCE, SOURCE, REMOVAL OF RESTRAINT, etc. Contextual information is vital to understand the semantic role and relations among concepts and thus identify the image schemas behind them. As a result, conceptual relations like TAKES_PLACE_IN (for processes) or LOCATED_AT (for entities) evoked by verbs like occur often point at the notion of EXISTENCE. The meaning of verbs is crucial when it comes to explore image schemas. The concordances in Table 4 suggest that verbs activate spatial relations during comprehension, as previous
123
Cogn Process
Fig. 2 Collocational patterns for the term pain
Table 4 Lexicalization of image schemas in defining contexts for pain
The most common reason why angina pain occurs is the reduced flow of blood to
EXISTENCE
treatment for that soreness. If sciatica pain occurs during pregnancy after that correct
EXISTENCE
able to better unwind. Nevertheless, if pain persists for many days then its much better succession, but then it’s gone. The longer pains usually persist for at least a day, up
COUNTERFORCE COUNTERFORCE
is always described as sudden and sharp pain originating in the abdominal cavity. The
SOURCE/BEGIN OF PATH
. It has been demonstrated with headache pain arising from the neck and from the facial
SOURCE/BEGIN OF PATH
recommended for cases in which gallbladder pain is severe enough to have prescription or
SCALE
it just wasn’t firing properly, and the pain was fairly intense , radiating down my
SCALE
the usual cause of back pain. This kind of pain is generally mild and relatively short-lived
SCALE
, may also sometimes be a neuralgia, the pain radiating down the inner side of the thigh hypertension who presents with abdominal pain radiating to the back. He’s going to need dissipates within a couple of months. When the pain radiates from a central location, such
VERTICALITY FRONT-BACK CENTRE-PERIPHERY
underwent surgery to relieve chronic back pain , but the procedure paralyzed him from the
REMOVAL OF RESTRAINT
sleep and fluid diet can easily ease the pain and symptoms of Diverticulitis. If you
REMOVAL OF RESTRAINT
problems, and provides treatments to alleviate pain , lack of mobility, and treat disease conditions REMOVAL OF RESTRAINT
123
fight the disease. The lesions heal, the pain subsides within 3 to 5 weeks and in most
REMOVAL
began the program. Within a few days the pain had receded significantly and within two
REMOVAL
disease. It's generally time-limited. Acute pain generally vanishes once the injuries heals
REMOVAL
Cogn Process Table 5 Lexicalization of image schemas in the proposition of PAIN[Vradiate ? prep] SPACE
CONTAINMENT
PATH
LEFT-RIGHT CENTRE-PERIPHERY CENTRE-PERIPHERY UP-DOWN FRONT-BACK NEAR-FAR NEAR-FAR CONTACT CONTAINER IN-OUT IN-OUT SURFACE SOURCE-GOAL
right-sided neck pain. The patient experienced to cold, with intense anxiety. Coldness It has contraction of the chest with the next time, with only groin pain. When menstrual cramps or gas pains? lower back after your current incident or maybe throat of Herpes B infection at a wound include of neck and radiating to upper limbs or for patients who complain of right-sided pain in the thoracic spine and clearly only a few things into considerable zone. The patient subsequently the buttocks and thighs
research by Richardson et al. (2003) showed. Such spatial elements ‘‘could be part of the metaphoric understanding that underlies much of our language, and is rooted in embodied experiences and cultural influences’’ (ibı´d.: 768). This, in turn, provides evidence for the perceptual-motor character of linguistic representations. For instance, the verb to radiate (with rel. freq. in corpus 7.11) implies literal spatial relationships, and is therefore a good example to illustrate the interaction between linguistic conceptual representations with perceptual–spatial processes. As shown in Table 5, not only does the proposition PAIN[Vmotion ? prep] activate spatial image schemas [LEFT–RIGHT, CENTRE-PERIPHERY, UP-DOWN], but also CONTAINMENT and PATH image schemas. The depiction of image schemas in ostensive definitions Similarly to how language encodes image schemas through the lexicalization of embodied experiences, images, insofar as pictorial representations of concepts, constitute the depiction of such knowledge about the world. They are alleged to be demonstrative definitions which convey in a visual format the meaning verbally transmitted through terms, and may supplement, if not replace, conventional definitions. If image schemas underlie some conceptual designations, like intensional definitions and defining contexts, there is no reason to think otherwise in the case of visual representations of specialized concepts or ostensive definitions. By studying the depiction of the image schemas lexicalized in the definitions and contexts of different types of pain, we would guarantee internal coherence between those data categories and ostensive definitions in terminological databases. Then, the degree of representativeness of images appears to be somehow linked to how explicit image schemas are depicted and how easy or difficult is for the viewer to recognize (either consciously or unconsciously) the image schema which lies behind.
pain radiating pain radiating pains radiating pain radiating pain radiating pain radiating pain radiating pain radiating pain radiating pain radiating pain radiating pain radiating pain radiating
to the right periscapular region from head in all directions, duration. from the heart in all directions down a leg, the almost automatic around to the front and back around the arms and legs. Contact away from the bite wound or blisters around chest wall. • Forward through the back, pains radiating into the chest out into other areas of the disc over the right shoulder. A CT scan from the buttock to the foot
In Fig. 3, image a represents the concept SCIATICA by showing the internal anatomy of the sciatic nerve from the lower back to the foot. The irritated nerve is coloured in red and the painful sensation is depicted by arrows showing the direction to which pain radiates. The pictorial representation of sciatica suggests a VERTICALITY image schema [UP-DOWN]. Image B, however, uses concentric circles to indicate that the intensity of LOW BACK PAIN decreases as it goes away from its source, but it seems not to spread. This is the main difference between sciatica and low back pain, as we mentioned above, at least from a sensory-perceptual perspective. The same happens with the depictions of INFARCTION and ANGINA in images C and D, where colour red has conventionally been used to represent painful areas during a cardiac episode. Image C represents the left arm, left part of the chest, neck, jaw and ear to which infarction pain typically spreads. Such a depiction points to a SPACE image schema [CENTRE-PERIPHERY], which is helpful to distinguish myocardial infarction from angina pectoris. Image D represents the chest cavity within which pain is felt during angina. Lighter red is used to depict occasionally referred discomfort in the upper abdomen, shoulders and arms. This pictorial representation evokes a CONTAINMENT image schema, according to which the CONTENT is pain and the CONTAINER is the chest. For the purpose of SKV, the most suitable images for their inclusion in terminological databases would be those which represent the embodied characteristics of the concept depicted, thus facilitating the convergence of the information visually provided by ostensive definitions and the linguistic information provided by intensional definitions and defining contexts. Lexicalization of image schemas in naming concepts We have thus far illustrated how image schemas and mental images shape the way we refer to concepts. These
123
Cogn Process
Concept Image schema
Visual representations
Image schema
SCIATICA
LOW BACK PAIN
(a) UP-DOWN
(b) NEAR
Source: Unknown author http://azpaincenters.wordpress.com/2010/11/05/sid elined-by-low-back-pain/
Source:http://www.ezgaitbrace.com/medicalbracea pplications.html
INFARCTION
ANGINA
(c) CENTREPERIPHERY
(d) IN-OUT
Source: http://www.doctortipster.com/3134myocardial-infarction-causes-risk-factors-signsand-symtoms.html
Visual representations
Source: J. Heuser [GFDL (http://www.gnu.org/copyleft/fdl.html) or CC-BYSA-3.0 (http://creativecommons.org/licenses/bysa/3.0/)], via Wikimedia Commons.
Fig. 3 Visualization of image schemas in the pictorial representation of
SCIATICA, LOW BACK PAIN, INFARCTION
cognitive tools influence the way we perceive and thus conceptualize knowledge. In terminology, the different ways of classifying a reality are called dimensions and concepts are often said to be multidimensional in that they can be named according to the different foci, a result of situatedness and embodiment typical of how we interact with the world, imbued in mental images and image schemas. In VariMed,5 a lexical database containing the different lexicalizations of medical concepts and their motivation, a module has been inserted to include the different dimensions of lexicalizations, i.e. the different ways of conceptualizing particular medical conditions, their signs and symptoms. VariMed database has a twofold purpose: (a) to help communication specialists to choose the most
appropriate term among the different lexicalizations of a specific medical concept, in a given medical context for a specific audience and (b) to serve as a lexical repository of medical variants that researchers in fields such as lexical semantics or applied linguistics can use with different purposes. Dimensions are related to the different ways of seeing (Croft and Cruse 2004) a reality, underlying that ‘‘perceptually […] the same symbols may lead to different classification of the same object’’ (Przybyszewski 2010: 53). Dimensions are thus relations that are made explicit in lexicalizations. In our project, we have arrived to a closed inventory of relations to allow for the classification of diseases and their signs and symptoms: AFFECTS, AGENT/RESULT, VISUAL_ATTRIBUTE_OF, NON_ VISUAL_ATTRIBUTE_OF, MADE_OF, TEMPORAL_ STAGE, INTENSITY, RESEARCHER/DISCOVERER, LOCALIZATION, HAS_ORIGIN_IN. When it comes to defining perception-based concepts, it becomes critical to underline the concept’s most salient
5
VariMed is a terminological database in the field of medicine intended for both researchers and lay people. Its main purposes are research and the popularization of medicine and medical language. It can be accessed at http://varimed.ugr.es.
123
and
ANGINA
Cogn Process Fig. 4 Visually grasping a concept expressed with different image-schematic lexicalizations in English and Spanish
visual attributes; in contrast with other terminological resources, in VariMed images are being recorded to bring to the user’s attention the perceptual features of medical concepts. Specifically, physical symptoms are often best ‘‘visualized’’ when shown in an illustration or a picture, whereas other concepts can do without a visual or ostensive definition. Figure 4 depicts a corkscrew oesophagus, a condition in which uncoordinated contractions of the oesophagus do not propel food effectively to the stomach. The corkscrew mental image is a more elaborated version of the BLOCKAGE image schema. In other words, such image schema is lexicalized in the term corkscrew through a metaphorical extension which leads to a resemblance metaphor which recalls the form of a corkscrew as a VISUAL_ATTRIBUTE_OF. Furthermore, given the importance of situating concepts in a particular experiential frame, defining contexts are being incorporated to illustrate the different elements of the concept’s intension and their relevance in various communication frames. In the following screen, the dimension LOCALIZATION is chosen as the relevant dimension for the term ‘‘abdominal pain’’. Such dimension represents the CONTAINMENT image schema as defined in 4.1, where pain is the CONTENT and abdomen the CONTAINER. Figure 5 shows the series of dimensions which were identified after examining a large corpus of texts. These dimensions are evoked by different term variants on the basis of the different foci highlighted when naming concepts like abdominal pain.
FbT premises posit that terminographical knowledge bases should adjust knowledge activation to the different user and usage needs. As a consequence, it is crucial to acknowledge the conceptual facets which are reflected in definitions (intensional, contexts and ostensive) and the designation of concepts, in order to serve the purpose of user with diverse access profiles.
A study on the representativeness of image-schematic pictures: preliminary results Experimental research is being conducted within the VariMed project to find evidence about the representational value of image-schematic pictures in terminological databases. The scope of the study is to check whether images containing a depiction of the image schemas behind some medical concepts are considered the best pictorial representations by potential users of the VariMed database. It is also our aim to prove whether the degree of specialization of terms has an influence on the type of image deemed as the most representative; in other words, whether popular terms used by lay people to talk about the human body, health and diseases, and specialized terms used by healthcare providers trigger different types of images. For that purpose, we intend to carry out four experiments, the first of which is currently being implemented. We have surveyed English-speaking lay participants and asked them to choose the most representative image of a set of four. They were prompted with a short excerpt
123
Cogn Process Fig. 5 Inventory of dimensions made explicit in the lexicalization of concepts
containing either a highly specialized term or a more explanatory phrase for medical concepts. In the following experiments, we plan to replicate the questionnaire and survey English-speaking healthcare providers. In the second stage, we will survey Spanishspeaking people with no prior medical knowledge and Spanish healthcare providers. This way we will examine the universal nature of image schemas or the potential cultural and linguistic binds (or biases) underlying conceptualization. We hereby present some preliminary results of the first experiment in which we surveyed English-speaking lay people about the most representative images depicting a series of register-based term variants (otalgia-earache, mastectomy-surgical removal of breasts, cephalalgiaheadache, etc.). Given chance performance of 0.25, preliminary results show a statistically significant preference (38.23 %) for image-schematic pictorial representations to the detriment of arbitrarily selected images, which had been chosen at random provided they bore some resemblance with the concept under scrutiny. However, no relevant difference was found in terms of representativeness between imageschematic representations of concepts designated by specialized formal terms (37.20 %), and image-schematic representations of concepts designated by informal popular terms (39.26 %). These figures point to the fact that the depiction of image schemas in medical images contributes
123
to the appropriate pictorial representation of medical concepts, whereas register does not seem to influence to any extent the representativeness of images. Nevertheless, further research is still needed to confirm these findings and be able to compare them with those obtained from healthcare providers in both Spanish and English-speaking contexts.
Concluding remarks We set out to describe how image schemas influence lexicalization of concepts and their definitions, either intensional, ostensive or in the form of contexts. Illustrations, as conceived by FbT, and its implementation in the VariMed database have been used, after a principled selection based on the premises of the SKV approach. Image-schematic illustrations, congruent with the image schemas underlying linguistic representations, seem to be most representative of the concept depicted to the detriment of those images which represent, neither explicitly nor implicitly, the conceptualization of medical knowledge. However, more systematic studies on image representativeness are still needed. Further research will include testing the familiarity of lexicalizations with a Likert-type scale and incorporating corpus-based contexts in all terminological data in order to enhance knowledge transfer and acquisition by user of terminological resources.
Cogn Process Acknowledgments This work was carried out while Dr Prieto Velasco was a visiting researcher at the University of Manchester with the support of a research Grant by the Spanish Ministry of Education (CAS12/00005) in the framework of the R&D project VARIMED (FFI2011-23120). I would like to express my gratitude to Dr Lynott and the colleagues of the Decision and Cognitive Sciences Research Centre and the Embodied Cognition Laboratory at the University of Manchester for their valuable and constructive suggestions to this research work.
References Auvray M, Myin E, Spence C (2010) The sensory-discriminative and affective-motivational aspects of pain. Touch Temp Pleas Pain Special Issue Neurosci Biobehav Rev 34:214–223 Aziz-Zadeh L, Casasanto D, Feldman J, Saxe R, Talmy L (2008) Discovering the conceptual primitives. In: Love BC, McRae K, Sloutsky VM (eds.) Proceedings of the 30th annual conference of the Cognitive Science Society 27–28. Cognitive Science Society, Austin, TX Barsalou LW (2003) Situated simulation in the human conceptual system. Lang Cogn Process 18:513–562 Barsalou LW (2008) Grounded cognition. Annu Rev Psychol 59:617–645 Barsalou LW, Santos A, Simmons WK, Wilson CD (2008) Language and simulation in conceptual processing. In: de Vega M, Glenberg A, Graesser AC (eds) Symbols, embodiment, and meaning. Oxford University Press, Oxford, pp 309–326 Connell L, Lynott D, Dreyer F (2012) A functional role for modalityspecific perceptual systems in conceptual representations. PLoS One 7:3 Croft W, Cruse A (2004) Cognitive linguistics. Cambridge University Press, Cambridge Cuenca MJ, Hilferty J (1999) Introduccio´n a la lingu¨´ıstica cognitiva. Ariel, Barcelona Evans V, Green M (2006) Cognitive linguistics: an introduction. Edinburgh University Press, Edinburgh Faber P (ed) (2012) A cognitive linguistics view of terminology and specialized language. De Gruyter Mouton, New York Galinski C, Picht H (1997) Graphic and other semiotic forms of knowledge representation in terminology management. In: Wright SE, Budin G (eds) Handbook of terminology management, vol 1. John Benjamins, Amsterdam, pp 42–61 Goldman A, De Vignemont F (2009) Is social cognition embodied? Trends Cogn Sci 13(4):154–159 Gottlieb H (2005) Multidimensional translation: semantics turned semiotics. In: Gerzymisch-Arbogast H, Nauert S Proceedings EU high level scientific conferences, Marie Curie Euroconferences MuTra: challenges of multidimensional translation. Saarbru¨cken 2–6 May 2005 Saarland Museum Modern Gallery, Saarbru¨cken Hostetter AB, Alibali M (2008) Visible embodiment: gestures as simulated action. Psychon B Rev 15(3):495–514 Huang X, Huang H, Liao B, Xu C (2013) An ontology-based approach to metaphor cognitive computation. Mind Mach 23:105–121 ISO 10241-1 (2011) Terminological entries in standards. Part 1: general requirements and examples of presentation. International Organization for Standardization, Geneve ISO 704 (2000) Terminology work: principles and methods. International Organization for Standardization, Geneve ISO 704 (2009) Terminology work: principles and methods. International Organization for Standardization, Geneve
Johnson M (1987) The body in the mind: the bodily basis of meaning, imagination, and reason. University of Chicago Press, Chicago Kilgarriff A, Rychly P, Smrz P, Tugwell D (2004) The Sketch Engine. Proc EURALEX 2004, Lorient, France; 105–116. http://www. sketchengine.co.uk Kress G (2010) Multimodality: a social semiotic approach to contemporary communication. Routledge, London Lakoff G (1987) Women, fire, and dangerous things: what categories reveal about the mind. University of Chicago Press, Chicago Louwerse M, Connell L (2011) A taste of words: linguistic context and perceptual simulation predict the modality of words. Cognitive Sci 35(2):381–398 Louwerse M, Jeuniaux P (2008) Language comprehension is both embodied and symbolic. In: de Vega M, Glenberg A, Graesser AC (eds) Symbols, embodiment, and meaning. Oxford University Press, Oxford, pp 309–326 Louwerse M, Jeuniaux P (2010) The linguistic and embodied nature of conceptual processing. Cognition 114:96–104 Louwerse M, Zwaan R (2009) Language encodes geographical information. Cognitive Sci 33:51–73 Louwerse M, Cai Z, Hu X, Ventura M, Jeuniaux P (2006) Cognitively inspired natural language based knowledge representations: further explorations of Latent semantic analysis. Int J Artif Intell Tools 15(1021):1–20 Lynott D, Connell L (2010) Embodied conceptual combination. Front Psychol 1(212):1–14 Mandler JM (2004) The foundations of mind: origins of conceptual thought. Oxford University Press, San Diego Melzack R, Casey KL (1968) Sensory, motivational and central control determinants of chronic pain: a new conceptual model. In: Kenshalo DR (ed) The skin senses: Proceedings of the first international symposium on the skin senses. Florida State University in Tallahassee, Florida O’Halloran KL, Smith BA (eds) (2011) Multimodal studies: exploring issues and domains. Routledge, New York Pickering MJ, Garrod S (2009) Prediction and embodiment in dialogue. Eur J Soc Psychol 39:1162–1168 Pollatos O, Fu¨sto¨s J, Critchley HD (2012) On the generalised embodiment of pain: how interoceptive sensitivity modulates cutaneous pain perception. Pain 153:1680–1686 Prieto Velasco JA (2008) Informacio´n gra´fica y grados de especialidad en el discurso cientı´fico-te´cnico: un estudio de corpus. Dissertation, University of Granada Prieto Velasco JA (2013) A corpus-based approach to the multimodal analysis of specialized knowledge. Lang Resour Eval 47(2): 399–423 Prieto Velasco JA, Faber P (2012) Graphical information. In: Faber P (ed) A cognitive linguistics view of terminology and specialized language. De Gruyter Mouton, Boston, pp 225–248 Prieto Velasco JA, Lo´pez Rodrı´guez CI (2009) Managing graphic information in terminological knowledge bases. Terminology 15(2):179–213 Przybyszewski AW (2010) Logical rules of visual brain: from anatomy through neurophysiology to cognition. Cogn Syst Res 11(1):53–66 Richardson DC, Spivey MJ, Barsalou LW, MCRae K (2003) Spatial representations activated during real time comprehension of verbs. Cogn Sci 27:767–780 Rohrer T (2009) Image schemata in the brain. In: Hampe B (ed) From perception to meaning: image schemas in cognitive linguistics. Mouton De Gruyer, NewYork Talmy L (1983) How language structures space. In: Herbert L, Pick Jr, Acredolo L (eds) Spatial orientation: theory, research, and application, Plenum Pres, New York Talmy L (2003) The representation of spatial structure in spoken and signed language. In: Emmor K (ed) Perspectives on classifier
123
Cogn Process constructions in Sign language. Lawrence Erlbaum, Mahwah, NJ, pp 69–195 Temmerman R (2000) Towards new ways in terminology description. The sociocognitive approach. Amsterdam, John Benjamins Temmerman R, Kerrremans K (2003) Termontography: ontology building and the sociocognitive approach to terminology description. In: Hajicˇova´, E, Koteˇsˇovcova´ A, Mı´rovsky´ J. (eds) Proceedings of CIL17, Matfyzpress, MFF UK (CD-ROM). Prague, Czech Republic Tercedor Sa´nchez M (2011) The cognitive dynamics of terminological variation. Terminology 17(2):181–197
123
Tercedor Sa´nchez M, Uren˜a Go´mez-Moreno JM, Prieto Velasco JA (2012) Metaphoric and metonymic processes in terminology. J Spec Transl 18:187–205 Williams LW, Huang JY, Bargh JA (2009) The scaffolded mind: higher mental processes are grounded in early experience of the physical world. Eur J Soc Psychol 39:1257–1267 Zwaan RA (2009) Mental simulation in language comprehension and social cognition. Eur J Soc Psychol 39:1142–1150 Zwaan RA, Stanfield RA, Yaxley RH (2002) Language comprehenders mentally represent the shape of objects. Psychol Sci 13:168–171
