Iberica 13 Ibérica 34 (2017): 17-44 ISSN: 1139-7241 / e-ISSN: 2340-2784 Abstract After briefly reviewing the cognitive-linguistic notions of metonymy and constructional form adhered to by the author (A) and discussing the general grammatical notion of clipping and A’s notion of “natural metonymic clipping”, the paper presents the list of salience factors whose combination determines the overall relative salience of a word segment. Two well-known inventories of American English medical abbreviations are then analyzed with the goal of identifying natural metonymic clippings in this register, noting their scarcity. A sample of three word segments that have become conventional medical clippings (tab- for tablet, -lytes for Electrolytes, and Chem panel for Chemistry panel) and a segment of one of the original full forms that has not become a conventional clipping (-blet in tablet) is then selected with the purpose of testing A’s salience factor grid on it. This grid is carefully described, including its numerical values, and systematically applied to the above-mentioned sample. The application of the grid to the sample seems to explain to a large extent the selection of the segments conventionalized as clippings in the sample, especially if compared to other “rival” segments. These results seem to confirm A’s earlier work arguing for the validity of the salience factor grid as a tool to account for the overall relative salience of a word segment and its (non) conventionalization as a natural metonymic clipping. Keywords: lexical morphology, metonymy, linguistic motivation, medical discourse, clipping. Resumen Fa c to r e s de s al i e nc ia d e te r m in an te s de l as ab r e v ia c i on e s me t on ími c a s n at ur al e s e j e mpl if i c ad a s m e di an te e l l é x ic o mé di c o Salience factors determining natural metonymic clippings illustrated through the medical lexicon Antonio Barcelona Sánchez Universidad Córdoba (España) antonio.barcelona@uco.es 17 Ibérica 34 (2017): 17-44 ANTONIO BARCELONA SÁNCHEZ Tras repasar brevemente las nociones cognitivistas de metonimia y forma construccional que propone el autor (A) y comentar la noción gramatical general de “clipping” (o “recorte” morfológico) y el concepto de “recorte metonímico natural” propuesto por A, el trabajo presenta la lista de factores de saliencia cuya combinación determina la saliencia relativa global de los segmentos de una palabra. A continuación se analizan dos conocidos inventarios de abreviaciones médicas en inglés americano con objeto de identificar recortes metonímicos naturales en ese registro, y se observa su escasez en el mismo. Se selecciona luego una muestra de tres segmentos léxicos que han pasado a ser recortes (clippings) convencionales en el leguaje medico (tab- por tablet, -lytes por Electrolytes, y Chem panel por Chemistry panel), y un segmento de una de las formas completas originales que no se ha convertido en un clipping convencional (-blet en tablet), con objeto de poner a prueba la plantilla de factores de saliencia propuesta por A. Se describe detalladamente dicha plantilla con sus valores numéricos, y se la aplica sistemáticamente a la muestra. La aplicación de la plantilla parece explicar en gran medida la selección de los segmentos de la muestra como clippings o recortes convencionales, sobre todo si se les compara con otros segmentos “rivales”. Estos resultados parecen confirmar investigaciones anteriores de A que apoyan la validez de la citada plantilla de factores de saliencia como herramienta para explicar la saliencia relativa global de un segmento léxico y su (no) convencionalización como recorte (clipping) metonímico natural. Palabras clave: morfología léxica, metonimia, motivación lingüística, discurso médico, recorte (clipping). 1. Introduction This article has been written with the goal of presenting to the academic community, especially the researchers in scientific and technical language, a salience factor grid that I have been developing over the last few years to account for the selection of the lexical segments eventually becoming standard abbreviated English lexical forms, known as “clippings”. Although this grid can be applied to explain in part all other types of lexical abbreviations and also syntactic abbreviations, I have so far only applied it to, and tested it with, what are called below “natural metonymic clippings” (Barcelona, 2016). In the present paper I continue testing the grid with a sample of conventional clippings drawn from the medical lexicon. The main contribution of this paper, as I see it, is thus not a systematic study of all types of medical abbreviations, most of which are not natural metonymic clippings (as will be explained below), but the evidence it provides of the 18 validity of that grid to account for all instances of natural metonymic clippings in all sorts of genres and registers. In section 2, I will briefly clarify the cognitive-linguistic notion of metonymy I adhere to and will argue for the claim that metonymy does not only motivate meaning construction but also certain types of “constructional form”. This latter term will also be explained in that section. In section 3, the notion of “natural metonymic” clipping will be very briefly characterized. This will be followed by a brief, initial enumeration of the list of salience factors involved in the selection of the segments retained as clippings, with particular attention to the “naturalness” and “ability to evoke the full form” factors. Section 4 will be devoted to discussing briefly the regularities observed in medical lexical abbreviations (one of them being the scarcity of natural metonymic abbreviations), and to presenting the small four-item sample of medical language clippings and the criteria followed in the selection. Section 5 is the main section of the article. It includes a more detailed description of the salience factor grid and its application to the sample of medical clippings selected in section 4. Section 6 includes the conclusions of the article. 2. Metonymy in cognitive linguistics and in constructional form The notion of “metonymic clippings” implies the notions of “metonymy” and “clipping”. The latter notion will be discussed in section 3. The present section is devoted to briefly discussing the notion of metonymy in cognitive linguistics (CL) and in constructional form. All cognitive linguists share the basic cognitive notion of metonymy, at least in its core elements. Everyone in the field agrees that, like metaphor, metonymy is a “conceptual” phenomenon whereby one concept (normally called the source domain) is used to activate another closely related concept. However, cognitive linguists disagree on a number of specific basic issues, such as the distinction between metonymy and related phenomena like “zone activation” and “facetization” (Paradis, 2011), the exact nature of the source-target link, the defeasibility of metonymy (Panther & Thornburg, in NATURAL METONYMIC CLIPPINGS ILLUSTRATED THROUGH THE MEDICAL LEXICON Ibérica 34 (2017): 17-44 19 press), the reality of WHOLE FOR PART and PART FOR PART metonymies, its distinction from metaphor and other phenomena (Barnden, 2010), and other issues, such as metonymic prototypicality (Barcelona, 2003, 2011; Peirsman & Geeraerts, 2006). Barcelona (2011), Kövecses and Radden (1998), Lakoff and Turner (1989), Panther and Thornburg (2003, 2007, in press), Radden (in press), and Bierwiaczonek (2013), among others, have discussed all these issues in detail. The review of their respective positions would go beyond the bounds of this article. The following definition of metonymy (adapted from Barcelona, 2011 and included in Barcelona, 2015) is an attempt at providing a general, “schematic” definition of metonymy: Metonymy is an asymmetric mapping of a conceptual entity, the source, onto another conceptual entity, the target. Source and target are in the same frame and their roles are linked by a pragmatic function, so that the target is mentally activated. This definition is not very different from other well-known definitions within CL provided by some of the above-mentioned linguists (especially Kövecses & Radden, 1998, and Panther & Thornburg, 2007), and should be regarded as a synthesis of them all, with some additional ingredients. The “mapping” or “conceptual projection” (Lakoff & Turner, 1989: 103-104) in metonymy occurs between entities in the same conceptual domain, whereas metaphor is a mapping occurring across two different conceptual domains. Whereas in metaphor we find a symmetric and systematic correspondence and structural equivalence between elements of the conceptual structure of the source and those of the target (Lakoff, 1993), in metonymy we do not find any degree of structural similarity or equivalence between source and target, hence their mapping is “asymmetric”. In the metaphor LIFE IS A JOURNEY, the beginning of the journey is mapped onto the beginning of life, the obstacles in the journey onto life’s difficulties, etc. But in a PART FOR WHOLE metonymy like BODY PART FOR PERSON as in (1) The ship was lost with all hands the source body part does not map symmetrically onto the whole person: the fingers in the hand are not projected onto any specific aspect of the person, the palm is not projected either onto any specific aspect of the person, and so forth. Only the hand (and the associated knowledge about its use) is used to activate the notion of a (specific type of) person (a passenger or a manual worker, a sailor in this case). ANTONIO BARCELONA SÁNCHEZ Ibérica 34 (2017): 17-4420 The term “frame” is probably more adequate than the term “domain”, used by Lakoff and Turner (1989: 103), to designate the conceptual structure within which metonymies operate. “Frame” designates a knowledge-rich model of a recurrent, well-delineated area of experience (Fillmore, 1985). Frames are equivalent to one of the types of Lakoff’s (1987) “Idealized Cognitive Models” (ICMs), namely “propositional ICMs”. The terms “domain”, “ICM” and “frame” are often used interchangeably but “domain” should probably be distinguished from the other two (Radden & Dirven, 2007: 9-12), due to its ambiguity. It can be used both in a “taxonomic” sense, to designate the schematic classification and subclassification into taxonomies of broad areas of experience, such as PHYSICAL ENTITIES in general, and in a “functional” sense, to organize our detailed knowledge about specific areas of experience (the domains then are called “functional domains” by Barcelona 2002a, 2003, 2011); in this use “domain” is synonymous to “frame” and “(propositional) ICM”. For further arguments for the claim that metonymy operate within frames, see Barcelona (2002a, 2011). Frames assign a “role” to the mental entities populating them. In example (1), the relevant frame is the HUMAN BEING frame, which represents speakers’ detailed encyclopedic knowledge about human beings (their bodies and body parts, physiology, emotions, interaction, life cycle, etc.). The source and target roles must be linked by what Fauconnier (1997) calls a “pragmatic function” and Kövecses and Radden (1998) and Radden (in press) a “metonymic relationship”. A “pragmatic function” is a privileged conceptual link in our long-term memory between the roles of metonymic source and target within the corresponding frame: BODY-PART FOR PERSON, CAUSE-EFFECT, PRODUCER-PRODUCT, AGENT- ACTION, CONDITION-RESULT, AGENT-INSTRUMENT, THING- REPRESENTATION, SALIENT PART OF CONSTRUCTIONAL FORM FOR WHOLE CONSTRUCTIONAL FORM etc. (see below on this last metonymy). This privileged link is an essential condition for the metonymic mental activation of the target by the source. The nose and mouth are distinct conceptual units included in the English-culture HUMAN PERSON frame, but their roles are not connected by a pragmatic function and therefore they cannot act as a metonymic source for the other (Kövecses & Radden, 1998: 48-49). However, the pragmatic function connecting the role SALIENT BODY PART assigned to mouth, nose, eyes and other body parts with the WHOLE role assigned to the PERSON or the NATURAL METONYMIC CLIPPINGS ILLUSTRATED THROUGH THE MEDICAL LEXICON Ibérica 34 (2017): 17-44 21 HUMAN BODY allows these body parts to act as metonymic sources for people, as in He only earns four hundred pounds a month and, with five mouths to feed, he finds this very hard. There is wide agreement that metonymies are ubiquitous in language, thought and communication. As regards language, metonymy has been argued to operate in phonology, grammar and discourse (Barcelona 2002b, 2013, 2015; Radden 2005), and to constantly interact with other metonymies, metaphors and other conceptual structures (Barcelona, 2005; Ruiz de Mendoza, 2011).1 Grammar in CL is a structured inventory of “grammatical constructions” (a key concept in CL), which are conventional pairings of form (including phonemic and prosodic form) and meaning (including all sorts of pragmatic meaning) at all levels, from morphemes through lexemes, phrases and several types of structures with an indeterminate hierarchical level like idioms and other relatively fixed expressions, to clauses and sentences (Goldberg, 2006; Langacker, 2008: 161-214). Metonymy operates in all sorts of grammatical constructions (see the various chapters in Panther, Thornburg, & Barcelona, 2009), mainly by directly or indirectly motivating their meaning but also by motivating their form and their form- meaning connection. The metonymic motivation of (initially non-canonical) “constructional form” did not receive enough attention by early metonymy researchers in CL (except for brief remarks on the topic by Kövecses & Radden, 1998: 45-46 and Radden & Kövecses, 1999: 28, 36). But, more recently, Barcelona (e.g. 2002b, 2005, 2013, 2009, 2016), Radden (2005: 17) and, especially Bierwiaczonek (2007, 2013: Ch. 2) have recently begun to explore this phenomenon. Constructional forms (including lexical forms) are “models” stored in speakers’ minds, and as models, they, too, constitute conceptual units, irrespective of their actual phonetic realization. This allows us to view the set of basic forms2 of a given construction (i.e. a lexeme) as constituting a conceptual frame, within which metonymy can operate. Constructional forms partially motivated by metonymy include certain lexical abbreviations, or “clippings” (like prob from problem), and certain types of ellipsis, as in “Did you buy the tickets” “Yes, [I bought the tickets] yesterday”; the metonymy at work is SALIENT PART OF CONSTRUCTIONAL FORM FOR WHOLE FORM. ANTONIO BARCELONA SÁNCHEZ Ibérica 34 (2017): 17-4422 3. Clippings and “natural metonymic clippings” The grammatical notion of “clipping” is, in principle, quite simple: in Morphology, it designates an abbreviated lexical form, that is, a shortened form of a lexeme resulting from the retention of just one part of its full form. Typical examples include prof for professor (where only the initial part is retained), fridge for refrigerator (where only a middle part is retained), or bus for the now near obsolete form omnibus (where only the final part is retained). Of course, this calls for the need to determine when the abbreviation affects a lexeme or a phrasal construction. For example, one may wonder whether such a sequence as the standard construction in medical English Chemistry panel, “A comprehensive screening blood test that indicates the status of the liver, kidneys, and electrolytes” is a (compound) lexeme or a phrase. In terms of the standard definition of compounding, it should be regarded as a compound rather than a multiply variable noun phrase, given its relative fixedness and its highly specific meaning. Therefore, the standard abbreviation Chem panel of this compound in medical English should count as a clipping affecting one of the lexical morphemes (Chemistry) constituting the compound.3 Most of the medical clippings in the sample analyzed are, however, forms of mono-morphemic lexemes. The term “natural metonymic clipping” will be reserved for the abbreviated lexical forms that retain a real, i.e. “natural”, segment of the full lexical form and can best evoke the full form. In my view, it is these clippings that can be regarded as most clearly metonymic clippings as far as their form is concerned, so that an abbreviation like hanky for handkerchief is less natural a clipping as doc for doctor (see below on the “naturalness” salience factor). The selection of the segment of the full form of a lexeme retained as a conventional metonymic clipped form of that lexeme is motivated by the relative salience of that segment within the full form. In turn, that salience is determined, so we claim, by the number of salience factors exhibited by the segment in question and the degree to which they operate in it. On the basis of the analysis and observation of over two hundred English clippings representing a wide variety of types, I identified the following salience factors (Barcelona, in preparation; Barcelona, 2016): A. Higher-weighted factors: Initial position, Being relatively easy to recognize and remember, Being relatively easy to pronounce, Ability to evoke the full form, Ability to evoke the meaning of the full constructional form, NATURAL METONYMIC CLIPPINGS ILLUSTRATED THROUGH THE MEDICAL LEXICON Ibérica 34 (2017): 17-44 23 “Naturalness” of the segment, Length (the shorter, the more salient), “Energetic” effect, Formal distinctiveness B. Lower-weighted factors: Prosody: Primary stress, “Audibility”, Similarity to existing full or shortened forms of other constructions, Final position, Medial position. C. Other factors: Aesthetic preference, Entrenchment. Others may have to be added in special cases. These factors will be briefly discussed and illustrated with medical clippings in section 5. They are treated in greater detail in Barcelona (2016) and Barcelona (in preparation). But, in order to round off the notion of “natural metonymic clippings”, we need to anticipate the discussion of the higher- weighted factors “‘naturalness’ of the segment” and “ability to evoke the full form”. A metonymic abbreviated form should in principle constitute a “natural” segment of the full form. By “natural” I mean the degree to which the graphemes and phonemes of the segment selected as a clipped form mirror the same “continuous” sequence in one of the segments of the full form. The scores (see section 5) reached by each segment on this factor are inversely proportional to the number of phonemes/graphemes or prosodic features (especially stress) included by the segment which do not occur in exactly the same environment in the full form. Hence the abbreviated form gas constitutes a natural segment of gasoline, since its phonemes/graphemes mirror the same continuous sequence as in the segment gas- of the full form. But the nonexistent abbreviated forms *gs or *gsln would not be natural in our sense, since, although they do pick out some phonemic/graphemic parts of the full form, their graphemic/phonemic sequences do not mirror any continuous sequence in the full form. An intended abbreviated form of gasoline like gassy, though perfectly mirroring the continuity of the segment gas- (in speech, not in writing, where an additional has been inserted), would be less “natural” in our sense because it adds one phoneme/grapheme which is not present in any continuous sequence of the full form (on the other hand, gassy is the full form of another lexeme with the meanings “abounding in gas; of the nature of gas”; see 1989 OED edition, revised 2009 for CD-ROM version). A similar reasoning applies to the low naturalness of hanky for handkerchief. The similarity of the clipped form to the full form is also determined by the total number of phonemes/graphemes and/or syllables shared by the ANTONIO BARCELONA SÁNCHEZ Ibérica 34 (2017): 17-4424 clipped form with the full form, whether or not they occur in the same continuous sequence. This is the type of similarity contemplated in the factor “ability to evoke the full form”, which is the other factor that determines the metonymicity of a clipping. The minimum degree of similarity required for the eligibility of a segment as the source form in the SALIENT PART OF FORM FOR WHOLE FORM metonymy is the sharing of one syllable with the target full form. The established clipping prof for professor matches this minimum requirement and so does the nonexistent clipping *essor for professor (although the overall salience of prof in terms of all the other salience factors is much higher than that of *essor). The difference between the factors “ability to evoke the full form” and “naturalness” can be seen by examining the hypocoristic clipping Bob (