How specific wine tasting descriptors are?

Procedia - Social and Behavioral Sciences 198 ( 2015 ) 287 – 299 Available online at www.sciencedirect.com ScienceDirect 1877-0428 © 2015 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of Universidad de Valladolid, Facultad de Comercio. doi: 10.1016/j.sbspro.2015.07.447 7th International Conference on Corpus Linguistics: Current Work in Corpus Linguistics: Working with Traditionally-conceived Corpora and Beyond (CILC 2015) How specific wine tasting descriptors are? Belen López Arroyo*, Roda P. Roberts Universidad de Valladolid-ACTRES, Valladolid, 47011, Spain University of Ottawa-ACTRES, Ottawa, K1N 6N5 Abstract Although wine tasting notes are a specialized genre in the field of Oenology, with their own rhetoric and language, some of the elements describing the language of wine are by no means as specialized as that of most other specialized genres such as technical notes, auditor reports or abstracts. In this study, we will explore how wine literature attempts to analyze such adjectives used in wine tasting notes. We will then, based on a comparable corpus of 700 tasting notes per language, study a number of common wine descriptors in English and Spanish in the context of the nouns that they collocate. Our purpose is to determine their specificity or generality in their use and meaning. On the basis of their collocability, we will categorize the descriptors and analyze the meaning components of the descriptors falling in the most general category. © 2015 The Authors. Published by Elsevier Ltd. Peer-review under responsibility of Universidad de Valladolid, Facultad de Comercio. Keywords: specialized corpus; contrastive analysis; genre studies; descriptors; collocates. 1. Introduction Wine tasting notes constitute a specialized genre in the field of Oenology, with their own rhetoric and language. However, the language of wine tasting notes is by no means as specialized as that of most other specialized genres such as abstracts, technical sheets, reports…. Indeed, while there are dozens of terms used to describe and evaluate wines, there are only a limited number of words that are used exclusively or primarily for describing taste. Based on the wine literature of the 1970s, Lehrer has listed 238 wine descriptors, which are the “commonest words” (2009, 4) that she found attested. Examples are astringent, fresh, meaty, and spicy. But when one says that a * Corresponding author. Tel.: +34 983 184270. E-mail address: [email protected] © 2015 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of Universidad de Valladolid, Facultad de Comercio. 288 Belen López Arroyo and Roda P. Roberts / Procedia - Social and Behavioral Sciences 198 ( 2015 ) 287 – 299 wine fresh, what aspect of the wine is being referred to: its aroma or its taste? And what dimension of aroma or taste is being alluded to? In this study, after examining how wine literature attempts to analyze such adjectives used in wine tasting notes, we are going to study a number of common wine descriptors in English and Spanish in the context of the nouns that they collocate with in order to determine how specific or general these descriptors are in their use and meaning. 2. Literature on wine descriptors Many efforts have been made over the years, both by wine experts and by linguists, to categorize and organize the many adjectives used to describe wines. In 1972, in their Essai sur la Dégustation des Vins, Vedel et al. proposed a red wine structure, the so-called “triangle de Vedel”, which represents the balance between the three components to be assessed in tasting a red wine: acidity, astringency and sweetness. Pertinent adjectival descriptors are provided for each of the three components, allowing for identification of the qualities and weaknesses of a red wine on the basis of supposedly objective criteria. In the 1980s, Noble developed a system for the description of the smells of wines, which has become widely known as the Aroma Wheel. The Aroma Wheel organizes the descriptors of aroma into three categories based on their specificity. The most general descriptors, which are found in the inner circle, are adjectives like floral, fruity, woody and nutty. These general descriptors are subdivided and made more specific in the next circle: fruity, for example, is subdivided into citrus, berry, tree fruit, melon, tropical fruit, and cooked fruit. Finally, each of the more specific fields designated by the latter is further subdivided in the outer circle: citrus, for example, is divided into orange, grapefruit, lemon and lime. While the general descriptors are primarily adjectives, many of the more specific ones are nouns. Another descriptors wheel, this one for mouthfeel perceptions, was developed by Gawel et al. (2000). The mouthfeel wheel terminology was partially intended to be a starting point for sensory panels to allow the rating of the intensity of defined characteristics of a set of samples, to provide a profile of the mouthfeel characteristics. The innermost layer of the wheel divides the mouthfeel terms into two categories: feel and astringency. The second tier presents groups of terms that are subordinate to these categories, such as for instance weight, harsh, surfacesmoothness. In the outermost tier, these categories are further subdivided. The terms viscous, full, thin, watery are thereby categorized as descriptors of weight, hard, aggressive, abrasive as descriptors of harsh, and furry, fine emery, velvet, suede, silk, chamois, satin as descriptors of surface-smoothness. The descriptors, which are primarily adjectives, are grouped as terms along a scale in the case of weight, although other groupings of terms (e.g. those describing surface-smoothness) are less relatable to points on a scale. Each of the terms on the wheel are defined, either by written definitions or by use of finger touch standards, so that each taster knows what is meant by the term. The British organization Wine and Spirit Education Trust has proposed a wine tasting template that provides wine descriptors organized in scales for the following properties of the wine (in Herdenstam 2004, 58): • Appearance • Intensity: pale – medium – deep – opaque • Color: purple – ruby – garnet etc. • Nose • Condition: clean – unclean • Intensity: weak – medium – pronounced • Development: youthful – grape aromas – aged bouquet – tired – oxidized • Fruit character: fruity, floral, vegetal, spicy, woods, smoky, animals etc. • Palate • Sweetness: dry – off-dry – medium dry – medium sweet – sweet – luscious • Acidity: flabby – low – balanced – crisp – acidic • Tannin: astringent – hard – balanced – soft • Body: thin – light – medium – full – heavy • Fruit intensity: weak – medium – pronounced 289 Belen López Arroyo and Roda P. Roberts / Procedia - Social and Behavioral Sciences 198 ( 2015 ) 287 – 299 • Fruit character: (same as for nose) • Alcohol: light – medium – high • Length: short – medium – long The classifications of descriptors discussed above have been proposed primarily by wine experts. However, more recently, linguists have also made efforts to categorize wine descriptors. Coutier (1994, 667), in her discussion of metaphors in the vocabulary of wine tasting, classifies metaphorical descriptors (and other wine tasting terms) according to the thematic fields from which the metaphors are drawn: • The human being (physical): charnu, musclé • The human being (mental): aimable, expansif, réservé, sincère • The human being (social): aristocratique, noble, racé • Spatial reality (form, volume, dimension): ample, cônique, rond • Physical reality (physical properties, state of matter): dense, solide, souple • Contact, movement, displacement: fuyant • Time (age, duration, evolution): court, sénile • The senses: asséchant, dur, ferme, frais, lisse • Textile : en dentelles • Construction : bien/mal construit Amoraritei (2002, 11) pursues essentially the same type of analysis, concentrating particularly on wine tasting descriptors related metaphorically to human beings - both their physical condition (e.g. charnu, charpenté, maigre, informe) and their psychological state (e.g. aimable, franc, généreux). A different type of analysis has been undertaken by Normand (1998). Her study of wine descriptors, which is corpus based and uses a variety of computerized tools, first identifies all the adjectives in her corpus, then classifies these adjectives according to the words which they qualify. If the qualifying and qualified words form a unit reflecting a perceptual judgement, the unit is retained and submitted to further analysis. The qualifying words are then organized into lexical classes in terms of the words they qualify, and presented in the form of a classificatory tree. This then allows the researcher to propose a structured representation of the adjectives, based on Rastier’s linguistic theory of interpretive semantics. The lexicon is structured on three levels: • Level I: Wine tasting phase (appearance, aroma, taste) • Level II: Property that is qualified (e.g. effervescence) • Level III: Quality/qualities related to the property Using this structure, Normand presents a representation of the qualities of the property robe du vin in the appearance phase: using the four quality classes of aspect, colour, intensity of colour and reflection, the researcher situates the descriptors limpide, trouble, lumineux, brillant, terne (all placed under aspect); jaune, or, doré, ambré (placed under colour); pâle, clair, soutenu, intense (placed under intensity of colour), and finally doré and vert (placed under reflection). Table 1. Lehrer’s dimensions of wine. DIMENSION TOO MUCH OF DIMENSION RIGHT AMOUNT OF DIMENSION TOO LITTLE OF DIMENSION Acidity sour tart flat Sweetness syrupy sweet N/A Balance unbalanced balanced unbalanced Astringency (Feel) harsh smooth N/A Age immature ripe withered Body coarse full-bodied weak Finish N/A lingering short The three dimensions mentioned by Lehrer have not been covered here. They are nose, activity and quality. That is because we are interested primarily in the Taste phase of the tasting note, while nose is part of the Aroma phase and Activity is part of Appearance. Quality is a dimension applied overall to the wine and is too general to be of interest to us. 290 Belen López Arroyo and Roda P. Roberts / Procedia - Social and Behavioral Sciences 198 ( 2015 ) 287 – 299 Lehrer’s study shows that a few adjectives appear on more than one dimension. She cites the example of sour, which is used to describe both acidity and balance, for balance is a function of both sugar and acid (2009, 28). She also discusses the descriptor flat, which occasionally means “too old” for wines intended to be drunk young or for sparkling wines that have lost their bubbles (2009, 7), but can also be used to describe a wine that has never had enough acid and therefore is flat from the beginning (2009, 7): “The wine words can be analyzed in terms of various dimensions. (…). A few words refer to complex properties and therefore appear on more than one dimension. Other words are ambiguous and appear more than once. For instance, flat can mean ‘lacking of acid’ or ‘an effervescent wine that has lost its bubbles”. While the distinction that Lehrer seems to be making between descriptors referring to complex properties and ambiguous descriptors remains vague, her reference to descriptors that appear on more than one dimension is of particular interest to this study, since we intend to analyze descriptors that collocate with nouns typical of different dimensions. 3. Methodology 3.1.Corpus The starting point of our study was an ad-hoc domain-specific (Corpas and Seghiri, 2009: 78) comparable corpus of wine tasting notes in English and Spanish, which we developed in the context of a larger research project. This corpus was compiled using pragmatic text selection criteria: the wine tasting notes were chosen to ensure a representative sample of the language of expert members of the discourse community, in this case of expert wine writers. We also considered availability, which refers to the ease with which the texts constituting the corpus could be obtained. The wine tasting notes corpus includes 750 wine tasting notes in Spanish and 716 wine tasting notes in English, which amount to 54,545 and 55,339 words respectively. While the corpus is relatively small, it is more than adequate for a specialized corpus and meets Biber’s criterion of needing at least 20 samples per register of between 2,000 and 5,000 words for a corpus to be useful for a research project (1993, 254). Moreover, according to Flowerdew (2005, 329), a smaller corpus may be called for when rhetorical tagging, which can only be done semiautomatically, is used. An important feature of our corpus is that its rhetorical structure is tagged. We annotated the texts using rhetorical labels that help us set up the semantic units (moves and steps, according to Swales 1990, 2002) that constitute wine tasting notes in every language. Following Bondarko’s (1984) and Krzeszowski’s (1990) parameters, our methodology consists in describing the central and peripheral semantic units in each language, and their cross-linguistic juxtaposition and contrast in order to obtain the prototypical structure for wine tasting notes for the target discourse community. We were able to identify five different moves in the wine tasting notes with various steps. See Table 2 below for the rhetorical structure: Table 2. Moves and steps for wine tasting notes. Introductory remarks (IR) Appearance (AP) Colour hue and depth Clarity Viscosity Effervescence Aroma (AR) Fragance Intensity Development Taste (TA) Flavors Finish Astringency Mouthfeel Body Balance Concluding remarks (CR) 291 Belen López Arroyo and Roda P. Roberts / Procedia - Social and Behavioral Sciences 198 ( 2015 ) 287 – 299 The rhetorical labelling of the texts allowed us to identify wine words that were typical of the different moves and steps. 3.2. Identifying the key nouns and descriptors in the steps of the taste move To limit the scope of our study of wine descriptors, we restricted ourselves to one important move found in all wine tasting notes: Taste. We first identified the key nouns found in each of the steps of this move: flavors, finish, astringency, mouthfeel, body and balance. In most cases, the name of the step was the key noun: this was the case with flavors, finish, mouthfeel, body, and balance. In the case of astringency, however, the key nouns were different from the name of the step: tannins on the one hand, acidity on the other. Acidity and astringency are closely related, as Lehrer has shown (2009, 11); they were therefore grouped together in one single step for the rhetorical labelling of our corpus. However, in the context of this study, they are viewed as two distinct steps. Using the nouns flavors, finish, mouthfeel, body, balance, tannins and acidity as a starting point, we added synonyms of these nouns to our list: note (a synonym of flavors), structure (a synonym of body) and palate, midpalate, attack, entry and aftertaste (specific terms related to mouthfeel). We proceeded in the same way in Spanish as we did in English. Presented below in Table 3 are the lists of nouns used in both English and Spanish as key words designating the steps: Table 3. Steps and key nouns in English and Spanish. STEPS KEY NOUNS KEY NOUNS Acidity Acidity Acidez Astringency Tannin/s, Astringency Taninos, Astringencia Balance Balance Acidez Body Body, Structure Cuerpo, Estructura Flavors Flavor/s, Note/s Notas, Sabor, Gusto Finish Finish Final Mouthfeel Palate, Mid-palate, Aftertaste, Attack, entry Retrogusto, Posgusto, Entrada, Ataque, Boca Then, using a browser that presents concordance lines and statistics, we looked for these nouns in our corpus with the aim of identifying the descriptors that collocate with these nouns. A descriptor had to appear at least five times with one or more of these nouns to be retained past the initial stage. 3.3. Classification of the descriptors into categories Based on the process described above, the descriptors were first classified as applicable to given steps. The descriptors were then further reduced if they did not appear with the key nouns of a given step at least five times. They were then classified more globally according to their generality or specificity. 4. Results: Classification of descriptors according the steps they occur Presented below in Tables 4 and 5 are the descriptors, the steps in which they occur first in English, then in Spanish. 292 Belen López Arroyo and Roda P. Roberts / Procedia - Social and Behavioral Sciences 198 ( 2015 ) 287 – 299 Table 4. Descriptors, and steps and number in English. DESCRIPTORS OCCURRING IN THREE STEPS OR MORE DESCRIPTOR STEP Balanced Acidity Flavors Mouthfeel Full Body Mouthfeel Flavors Juicy Flavors Acidity Mouthfeel Astringency Rich Mouthfeel Flavors Body Astringency Finish Smooth Astringency Mouthfeel Finish Soft Astringency Mouthfeel Acidity Finish Well integrated Astringency Acidity Flavors DESCRIPTORS OCCURRING IN TWO STEPS DESCRIPTOR STEP Bright Flavors Acidity Crisp Acidity Flavors Fresh Flavors Acidity Lush Mouthfeel Flavors Silky Mouthfeel Astringency Supple Astringency Mouthfeel Sweet Flavors Acidity Toasty Flavors Finish Velvety Mouthfeel Astringency DESCRIPTORS OCCURRING IN ONE STEP 293 Belen López Arroyo and Roda P. Roberts / Procedia - Social and Behavioral Sciences 198 ( 2015 ) 287 – 299 Light Body Long Finish Medium Body Creamy Mouthfeel Ripe Flavors High Acidity Intense Flavors Lingering Finish Low Acidity Mineral Flavors Moderate Acidity Natural Acidity Oak Flavors Zesty Flavors Round Mouthfeel Luscious Flavors Spicy Flavors Floral Flavors Herbal Flavors Earthy Flavors Berry Flavors Table 5. Descriptors and steps in Spanish. DESCRIPTORS OCCURRING IN THREE STEPS OR MORE DESCRIPTOR STEP Fino/a(s) Astringency Mouthfeel Flavors Fresco/s/a(s) Acidity Mouthfeel Finish Flavors Frutal(es) Flavors Mouthfeel Finish Gran Body Flavors Acidity Mouthfeel Finish Maduro/s/a(s) Astringency Flavors Mouthfeel Sabroso(s) Mouthfeel Flavor Finish Suave(s) Mouthfeel Astringency Flavors 294 Belen López Arroyo and Roda P. Roberts / Procedia - Social and Behavioral Sciences 198 ( 2015 ) 287 – 299 DESCRIPTORS OCCURRING IN TWO STEPS DESCRIPTOR STEP Amable Mouthfeel Acidity Carnoso(s) Mouthfeel Body Elegante(s) Body Astringency Envolvente Mouthfeel Finish Equilibrado/s/a(s) Acidity Mouthfeel Especiado/s/a(s) Flavors Finish Integrado/s/a(s) Astringency Flavors Intenso/s/a(s) Flavor Mouthfeel Largo/a Mouthfeel Finish Ligero/s/a(s) Mouthfeel Flavors Persistente(s) Finish Mouthfeel Potente(s) Mouthfeel Body Redondo(s) Mouthfeel Astringency Sedoso(s) Mouthfeel Astringency Tostado/s/a(s) Flavors Mouthfeel DESCRIPTORS OCCURRING IN ONE STEP Amplio(s) Mouthfeel Aterciopelado(s) Mouthfeel Bajo Flavors Balsámico(s) Flavors Bien/Muy equilibrado Acidity Corpulento Body Dulce(s) Astringency Estructurado(s) Mouthfeel Expresivo(s) Mouthfeel Graso(s) Astringency Mouthfeel Medio Body Mineral(es) Flavors Natural Acidity Rico(s) Mouthfeel Seco(s) Mouthfeel Sutil(es) Mouthfeel Untuoso/a Mouthfeel 295 Belen López Arroyo and Roda P. Roberts / Procedia - Social and Behavioral Sciences 198 ( 2015 ) 287 – 299 Vivo Mouthfeel In total, there were 37 descriptors identified in English, and 40 in Spanish. The total number of descriptors per language is relatively similar, only 21 of the English descriptors apply to more than one step, while 22 of the Spanish descriptors do the same. Moreover, the maximum number of steps that the descriptors apply to does not vary depending on the language: five in both corpora. This classification of descriptors, based on their collocational possibilities with key nouns in different steps, reveals their generality or their specificity. Those descriptors that collocate with key nouns from three or more steps are obviously more general than those that collocate with key nouns from a single step. Rich, occurring in five steps versus Silky occurring in two in the English corpus; Fresco occurring in four steps versus Carnoso occurring in two in the Spanish corpus. The question that arises at this point is what makes certain descriptors more general in their application. Is their meaning more generic or do they cover two or more distinct meanings? That is what we will explore next. 5. Discussion: The most general descriptors and their meaning In this paper, as an example of our study, we will only show the results of those descriptors occurring in three or more steps. In an effort to determine what makes certain descriptors more general in their application in wine tasting notes than others, we followed a twofold procedure: (1) we looked up the general descriptors identified above in English and in Spanish in three monolingual wine glossaries/dictionaries to see if more than one sense or a generic sense could be identified therein; (2) we then analyzed the general descriptors in context, using our corpus, to see if that could provide any further meaning clues. We used the dictionaries/glossaries cited below for verification. It should be noted that wine dictionaries/glossaries are notoriously incomplete. Our choice was based on the appearance of at least a few of the descriptors we were studying in the glossary. English: http://www.zachys.com/oneoone/?id=54 http://www.winespectator.com/glossary/index/word/A http://www.napacabs.com/terminology.aspx Spanish: http://www.jmcaro.info/diccionario.htm#-Shttp://www.diccionariodelvino.com/index.php/tema/5/ http://www.delbuencomer.com.ar/index_archivos/diccionariodelvinoa.htm Of the seven general descriptors in English (those that applied to three or more steps), only four were found in the dictionaries consulted, in any form whatsoever: balanced, full, rich and soft. Balanced was found only in nominal form (Balance), and full was found as full-bodied. Juicy, smooth and well-integrated were not present in the dictionaries selected or any dictionary consulted. On the basis of the dictionary entries consulted, balanced and full can be considered as having a generic sense, which allows them to be applied to different steps: Acidity, Flavors, and Mouthfeel for balanced, and Body, Mouthfeel and Flavors for full. For example, according to http://www.winespectator.com/glossary/index/word/S. A wine is balanced when its elements are harmonious and no single element dominates. The "hard" components—acidity and tannins—balance the "soft" components—sweetness, fruit and alcohol. We consider this definition as all-encompassing, allowing for the descriptor balanced to be applied to several different aspects of wine. Two of the three definitions of soft clearly indicate that the same definition can be applied to either Astringency or Acidity. In other words, this entry for soft could have been written as follows: