Maria-Cornelia Wermuth
Instrumentality in medical classification rubrics[1]
Maria-CorneliaWermuth
LessiusUniversityCollege, Antwerp
1. Introduction
The aim of this paper is to expand on instrumentality in so-called medical classification rubrics [CR’s]. This investigation should not only provide some insight into the diversity of instrumentality, but also lead to some suggestions for applying cognitively inspired work in the domain of medical language studies.As such it furthers earlier research into the conceptual structure of CR’s(Wermuth 2005) which provided among others good evidence for the pivotal importance of instrumentalityin this highly specialized type of discourse.Starting from a cognitive conception of instrumentality as part of causality (Talmy 2000),I will provide an account of the different instrumental types and their linguistic encoding.The analysis used for this investigation is usage-based in the sense that I attend to a particular semantic component in this specific text type, viz. instrumentality in order to observe which formal elements in English[2] express this semantic component in different surgical domains (cf. also Gries & Stefanowitsch 2006).The examples are taken from different surgical disciplines and illustrate both the conceptual multi-layeredness of instrumentality and the linguistic potential of medical language to convey instrumental meaning.
For this exploration, first a number of observational and theoretical issues are needed, including an overview of the relevant features of CR’s, and a short review of
the cognitive approach to instrumentality and causality starting from Talmy’s causal model (Talmy 2000).
2. Background
2.1 Medical classifications
Given the specificity of the data we start with a brief description of medical classifications and the characteristics ofclassification rubrics [CR’s].
For financing reasons, the registration of medical data is of prime importance in health care. The majority of these data are established in natural language. Medical classification systems are used to convert these data into some structured format (i.e. numerical codes) in order to allow for their electronic processing. A large number of such systems are in use like The International Statistical Classification of Diseases and Related Health Problems (ICD), the Diagnostic and Statistical Manual of Mental Disorders (DSM) and the Systematized Nomenclature of Medicine (SNOMED), just to name a few examples. Classification systems serve a variety of applications like the statistical analysis of diseases and therapeutic actions. For the purpose of this investigation we used a corpus of 500 CR’s in English which have been drawn from the procedure volume of the ICD-9-CM (International Classification of Diseases, 9th revision, Clinical Modification, Hospital Edition)[3]. The CR’s under investigation are natural language descriptions of codes referring to procedures on the ear (cf. Figure 1), on the eyes and on the cardiovascular, digestive and nervous system.In the next section I will elaborate on a number of relevantfeatures of these text strings.
Figure 1. Codes and classification rubrics(ICD-9-CM, chapter 4)
2.2 Medical classification rubrics [CR’s]
2.2.1 Expert-to-expert discourse
CR’s servein the first place a pragmatic function, which isto predicate the meaning of numerical codes for human users. Thissupports the linkage of codes withthe corresponding medicaldatain written sources and thus their computer-based processing.CR’s may be seen as propositionsreferringto scenarios (cf. Langacker 1987: 63ff.)which virtually are ‘universal’ in view of the heavily conventionalized way in which medical objects (i.e. procedures in the given case) are conceptualized independent from socio-cultural contexts (concerning the mental processing of propositions at the conceptual level I refer to Fauconnier 1985).BecauseCR’s are designed by medical specialists for medical specialiststhey only provide those cues which are assumed to be adequate to access the appropriate procedure scenario and to reconstruct the intended meaning of the code. Theexpert-to-expert discourse triggered by CR’sbetween the designer and the addressee (the coder) necessitates of course a common ground of shared (specialized) knowledge in order to be effective[4].This ‘partial’ representation of procedures remembers Talmy’s (2000) statement that that what is incidential and what is essential in an event is relative (i.e. it depends on the situation) and is specified by the actual wording.With respect to CR’s, for example, we observe thataspects like the operative approach (e.g. percutaneous, invasive), the operated body part (e.g. mitral valve, stomach), the pathology (e.g. tumor), etc. are typically encoded because they are classification-relevant aspects of procedure events.By contrast, the subject specifying the initiator agent (i.e. the surgeon taking the initiative and control of the action) is not mentioned as this knowledge is an inherent part of the conceptual representation of the operation scenes evoked by CR’s and, in addition, irrelevant in classification terms.
2.2.2 Nominalization and reification
This implicitness of CR’s is reflected by their surface structures. In morpho-syntactic terms, CR’s are reduced phrasal forms of (complex) sentences resulting in nominal phrases and a sequence of prepositional phrases. In their most simple realization, CR’s consist of a nominalized action verb (e.g. transplantation, replacement, insertion) or a neoclassicial root form (e.g. -tomy, -plasty, -ectomy) which is complemented by a number of premodifications like the examples given in (1), (2) and (3).
(1) Hearttransplantation
(2) Closed heartvalvotomy
(3) Single vessel percutaneous transluminal coronaryangioplasty
The head may also be followed by an of-phrase which introduces the DO like in (4), and the DO may be followed by other PP’s like in (5) and (6):
(4) Intraoperativemanipulationofstomach
(5)Removal offoreign body from esophagus without incision
(6)Insertion of permanent tube into esophagus
This elliptical nominal structure is derived from syntactically deeper structures of a different form. This form is more closely reflected in complex surface sentences, which are for example used in the detailed proceduredescriptions in op-notes. Here, the active verbal description highlights the temporal sequence of the actions. In CR’s, by contrast, we see that the basically sequential mode of the referent scenes is presented from a synoptic perspectival mode. Instead of active verbs, action nominalizations are used to express the dynamic actions which operations basically are. Because of the fact that the causal sequence of the separate events is expressed as a unitary event, the inherent causal multi-layeredness is not inferable from the predicate-argument structure. In other words, in CR’s an act-to-object reconceptualization takes place which can be termed a conceptual reification. In the verbal construction the act is represented in terms of ‘Agent affecting Patient’, whereas its nominalization makes thatthe referent is reified and conceptualized as an object. Reified actions are amenable to manipulations typically associated with a physical object such as pluralization, modification and quantification like in the examples (7-9).
(7) Intraoperative manipulations of stomach (pluralization) (8) Endoscopic dilation of pylorus (modification)
(9) Total gastrectomy with intestinal interposition (quantification)
2.2.3 The event structure of CR’s and windowing
As nominal text strings CR’s referto what Talmy (2000/2: 213ff.) termsa macro-event. This macro-event consists of a main or framing event and a number of subordinate events which bear some type of relation to the main event like cause or enablement. Together with the different subordinate events the framing event forms a causal chain and also provides the ‘upshot’, that is the conceptual framework or the event schema of the referent procedure. It is noteworthythat this event structure cannot directly be derived from the predicate-argument structure of CR’s which are basically elliptical nominal phrase constructions (cf. subsection 2.2.2). This means that the main predicate is not necessarily identical to the main event, because the attention over the referent scene (being some kind of procedure) is directed in a certain type of pattern that Talmy (2000: 76) terms the ‘windowing of attention’. In fact, in CR’s one or more portions of the sequential referent scene is placed in the foreground, whilst an important conceptual portion of the remainder of the scene is backgrounded (e.g. by means of PP’s) or simply omitted. The linguistic devices for the setting of attentional salience are the head position of the action verb within the nominal construction. Examples with the windowed classification-relevant event are given in (10-12).
(10)Cardiotomy [subevent]
(11)Stapedectomy [main event] with incus [subevent]
(12)Removal [subevent] of gallstones from liver
As these examples show, the windowed event not necessarily is identical to the main event, like in the example (10): In fact, any incision of the heart (cardiotomy)isby definition a subsequent event enabling the performance of some not further specified main event such as the operation on a heart valve.
What is important is the fact that although only a certain portion of the referent scene is explicitly specified by linguistic means, it is understood as part of the nature of the windowing process that – given the appropriate context – the addressee will be able to infer the remainder of the scene and conceptualize the referred procedure in an appropriate way. This is plausible if we consider that CR’s are particular constructs which render a more or less conventionalized particular conceptualization of a scene in medical reality (cf. also subsection 2.2.1). We therefore may conclude that the windowed event triggers an event frame: a set of conceptual elements that co-evoke each other constitute an event frame, while the incidental elements lie outside this frame. So, for example, not included within an operation event frame will be the day of the week on which an event occurred or the body mass index of the patient participating in the surgical event, even though such factors are part of the event.
In the given context we briefly note that Talmy’s notion of event frame is very close to Fillmore’s (1982) concept of a frame or scene when applied to an event, but there appear to be conceptual differences. First, Talmy stresses the exclusion of unprivileged elements from the core elements (Fillmore emphasis the co-presence of interrelated conceptual elements). Second, Talmy’s event frame is understood as a generic category that is universal across languages and that corresponds to the
structuring in other cognitive systems such as visual perception (Fillmore ‘s frames are language-specific and determined within a particular socio-cultural context).
In the medical context, event frames obviously have a generic language-independent status.
2.2.4 Figure and Ground
Finally, there are some interesting observationsto be made concerning the Figure-and-Ground structure of CR’s (Talmy 2000: 184).
As known, in full complex sentences the earlier event is seen as the reference point or Ground, and the later event (as requiring referencing) is the Figure. The Ground is in the subordinate clause, the Figure is in the main clause. In CR’s the designated events also have a Figure-and-Ground structure: We see that the windowed NP may be either a Figure or a Ground concept. The following examples illustrate that the Figure and Ground elements may be lexicalized by means of a PP, but also as the main NP. It is noteworthy that both Figure and Ground represent concepts of different categories like actions, body parts or artefacts.For example, in (13) the surgeon carried out an implantation (Ground) of a pulsation balloon (Figure) which suggests that the NP Implant is the reference event (Ground), whereas the pulsation balloon represents the Figure. This Figure is an artefact which is moved from outside into the body in which it has an instrumental function. In other words, there is an instrumental relationship between the Ground or main event Implant and the Figure pulsation balloonwhich can be characterized as Instrumentality MeansArtefact (cf. subsection3.4).
(13) Implant [Ground] of pulsation balloon [Figure]
(14)Medical [Figure] induction of labour [Ground]
(15)Reconstruction of eyelid [Ground] with flaps or grafts [Figure]
From the above description we can conclude that CR’s arelinguistic constructions which present an event as autonomous without the explicit encoding of the causal process during its occurrence. In Talmy’s (2000: 457) terms, in CR’s causality and accordingly explicit instrumentality fall outside the scope of attention which implies that both may be inferred indirectly. In the following sections I focus more specifically on instrumentality and its specific conceptualization and linguistic representation in CR’s.
3. Instrumentality
3.1 Artefactual vs. embedded instrumentality
In traditional definitions the instrumental case (also called the eighth case) is defined as a grammatical case used to indicate that a noun is the instrument or means by or with which the subject achieves or accomplishes an action. The noun may be either a physical object or an abstract concept. In the same vein,WordNet(Fellbaum 1998) defines instrumentality as “an artefact, or a set of artefacts, that are instrumental (i.e. behave as instruments) in accomplishing some end” (i.e. reaching some goal).In view of the technical character of surgical procedures, one could expect a lot of artefactual instrumentality expressed by some prepositional phrase introduced with prototypical instrumental prepositions like ‘with’ or ‘by (means of)’. However, the apparently obvious instrumental reading of with-phrases in CR’s can be completely wrong as the example in (16) illustrates:
(16) Replacement of any type pacemaker device with dual chamber device
The interpretation of the PP ‘with’ in this rubric has two plausible readings of which only the second one is correct in the given medical context. The first reading starts from the assumption hat the PP with dual chamber device is an adjunction of the verbal phrase replacement which results in the instrumental reading that any pacemaker (i.e. not necessarily a two chamber pacemaker) which has been implanted at an earlier stage is replaced by means of a pacemaker device that supports the stimulation of the two heart chambers. In the second reading, which is the correct one, the PP is interpreted as a modifier of the NP pacemaker device. The rubric’s meaning then is that all types of two heart chamber devices are replaced. In this reading, the ‘instrument’ (viz. surgical devices)used to perform this operation are simply gapped.
However, although there is a low salience, PP’s introduced by ‘with’code instrumentality in some way or another like theexample in (17) illustrates:
(17) Reconstruction of eyelid with flaps or grafts
This rubric consists of a causative construction that refers to the final subevent, namely that the eyelid has been reconstructed by using (with) flaps and grafts. In other words, these flaps and grafts are a kind of ‘bodily’ instruments because they permanently substitute the defective eyelid after its reconstruction. Of course, again a lot of prototypical instruments (i.e. surgical devices like scissors, scalpels and catheters)are indispensable for the procedure’s performance, but they are gapped in the surface structure being part of common background knowledge associated with the referent scene. The same holds forother aspects like the initiatory agent, thebodily motions the agent undertook to execute the intention – for example, taking a scalpel and incising the eyelid. So, starting from the predicate-argument structure of (17), the flaps or graftscould be assigned a Patient or Theme role being the entities which are transferred and implanted. However, from a medical real-world perspective, the flaps or grafts are perceived and also conceptualized as genuine instrumental entities: They are body parts which both are used as and function as an instrument. I term this phenomenon ‘embedded’ instrumentality (cf. subsection 3.1) because the factual instrumentality of the flaps or grafts cannot directly be derived from the surface structures. Consequently, in contrast to Fillmore who associates the semantic roles according with fixed prepositions like from (source), to (goal), along (path) and with (instrument), we follow Talmy (2000: 339) who states that the Figure in the causing event in turn is the instrument in the whole causative situation. This approach allows demonstrating that instrumentality in CR’s can be expressed by a large number of different markers and other linguistic structures that add distinct meaning (e.g. determiners in compounds, adjectives and deverbal nominalizations like in the examples (18-20).
(18) Forceps rotation of fetal head (= by means of a forceps)
(19)Medical induction of labor (= by means of drugs)
(20)Catheterization of Eustachian tube (= by means of a catheter)
In fact, a thorough investigation of the corpus showed two things. First, contrary to expectations, the prototypical instrumental preposition ‘by means of’ or ‘with’(like in the well-know example I cut the bread with a knife)rather infrequently introduces the instrument.Second, the examination of my dataclearly indicates that in the given context an essentially ‘artefactual’ conception of instrumentality is much too restrictive. By contrast, instrumentality in CR’s displays a whole range of subtypes, including embedded instrumentality in which actions or causes function as instruments.
The next subsection describes how a cognitive account of instrumentality, in which instrumentality is embedded into the event structure, can be fruitfully applied in order to embrace the diversity of instrumentality to be encountered in CR’s.
3.2 A cognitive approach to instrumentality
My analysis of instrumentality follows Talmy’s (2000: 474) account of instrumentality which opposes a number of distinct causation types to the single situational notion of causation in traditional linguistic treatments. The following summary mentions the most basic issues relevant to my present purpose.
Talmy’s (2000: 487ff.) discussion of the semantics of causation starts from the idea that there is a basic causative situation which consists of three components: a simple event, a causing event, and the causal relation (‘result from’) between the two events. The caused event functions as the Figure and the causing event as the Ground of the whole situation. According to Talmy the Ground element of the causing event is also the object which functions as the Figure of the caused event. An essential prerequisite is the (initiated or maintained) continuous force-exertional contact of the Figure of the causing event with this object. Moreover, the objects that function as the
Figure-and-Ground elements in the caused event fulfill the same functions relative to the whole causative event. The Figure-object of the causing event is instrumental with respect to the entire causative situation. Talmy (2000: 494) further stresses the interdependence between caused and causing event: The caused event is the ‘vector resultant’ of a sum of ‘Figural and Instrumental vectors’, which means that the force exerted by the Instrument of the causing event overcomes the natural motion tendency of the Figure of the caused event.
3.3 The causative event structure of CR’s
In principle, CR’s can be seen as genuine causative expressions as any surgical procedure is caused by some pathology. They basically describe embedded causative events, which are temporally included and contingent on each other. According to Talmy (cf. supra) the instrument is embedded into the causing event: the caused event RESULTS FROM the causing event, where the causing event has the structure: INSTRUMENT ACTS ON OBJECT (the object being in some way related to the object in the caused event). Because instruments can be viewed as causes of an event, we can conclude that any surgical procedure is inherently instrumental. To give an example: The meaning of the rubric in (21) is that a lacrimal cyst is incised (e.g. by means of a scalpel) and the fluid within this cyst is drained by means of some unspecified instrument.