Access to wine experts' long-term memory to decipher an ill-defined sensory concept: The case of green red wine

OENO One 2021, 1, 69 - 79 69 © 2021 In e na ional Vi icul u e and Enology Socie y - IVES Recei ed: 19 Augus 2020 y Accep ed: 3 Decembe 2020 y Published: 11 Janua y 2021 DOI:10.20870/oeno-one.2021.55.1.4475 Access o wine expe s’ long- e m memo y o deciphe an ill-de ined senso y concep : he case o g een ed wine Ma ía P. Sáenz-Na ajas1*, Ignacio A ias-Pé ez2, Sa a Fe e o-del-Teso1, Ana Escude o2, Vicen e Fe ei a2, Pu i icación Fe nández-Zu bano1, Dominique Valen in3 1Ins i u o de Ciencias de la Vid y del Vino (CSIC-GR-UR). Depa men o Enology, Finca La G aje a, C a. de Bu gos Km. 6 (LO-20 - salida 13), 26007 Log oño, Spain 2Labo a o io de Análisis del A oma y Enología (LAAE), Depa men o Analy ical Chemis y, Uni e sidad de Za agoza, Ins i u o Ag oalimen a io de A agón (IA2) (UNIZAR-CITA), Associa e uni o Ins i u o de Ciencias de la Vid y del Vino (CSIC-GR-UR), c/ Ped o Ce buna 12, 50009 Za agoza, Spain 3Cen e des Sciences du Goû e de l’Alimen a ion, INRA, Uni e si é de Bou gogne F anche-Com é, Dijon, F ance *co esponding au ho : mpsaenz@ic .es abs ac The p esen s udy aims o unde s and an ill-de ined senso y concep by a long- e m memo y-based s a egy wi h Spanish winemake s om ou wine egions using “g een wine” as a case s udy. A o al o 77 Spanish winemake s om ou Spanish wine egions ca ied ou a non- as ing ee desc ip ion ask. The desc ip ion ask yielded e ms belonging o wo main ca ego ies including o igin- ela ed e ms as well as senso y e ms. Senso y e ms belonged o a oma, as e, igeminal, colou , mul imodal and hedonic subca ego ies, which elucida es he mul idimensionali y o he s udied concep . The mos ci ed speci ic e ms we e “ ege al a oma”, “bi e ” and “unpleasan ”. Despi e hese commonali ies, a ce ain idiosync asy linked o as e (“excessi e sou ness”) and igeminal (“as ingency”) subca ego ies as well as o wine componen s (“ annins”) was e idenced as hey we e ci ed dis inc ly by expe s belonging o sepa a e wine egions. The capaci y o app oaches based on long- e m memo y o deciphe mul idimensional and ill-de ined concep s is highligh ed. The egional e ec is also explained in e ms o cogni i e p ocesses (i.e., knowledge and expe ience), which is linked o he use o senso y concep s by wine expe s. keywo ds as ingency, annins, ege al a oma, a oma, la ou , op-down, bo om-up, cogni i e © 2021 In e na ional Vi icul u e and Enology Socie y - IVES 70 OENO One 2021, 1, 69 - 79 Ma ía P. Sáenz-Na ajas e al. INTRODUCTION The e alua ion o ood senso y cha ac e is ics, o senso y p o iling, p o ides aluable in o ma ion o he ood indus y. P oduce s ha e o know he exac senso y p o ile o hei p oduc s o quali y con ol pu poses and o ensu e ha he p oduc ion p ocess ul ils he p oduce ´s equi emen s. In e ms o p oduc de elopmen , senso y cha ac e is ic e alua ion can help in he p oduc ion p ocesses by de ining he ing edien composi ion and hus achie e a speci ic and desi ed senso y p o ile. In classical consume science, he linkage o a p oduc ´s senso y p o ile o he consume ´s p e e ence o pe cei ed quali y app aisal (i.e., hedonic a ings) is a e y common s a egy. I allows he iden i ica ion o he main senso y a ibu es which d i e he consume p e e ence o any single p oduc . Con en ional desc ip i e analysis (DA) is among he mos widely used echniques o senso y p o iling. DA equi es a panel which is usually composed o a minimum o eigh pa icipan s. In each p oduc , he panel mus ha e p e ious expe ience wi h he speci ic senso y a ibu es o be e alua ed. The selec ed a ibu es ha e o be associa ed wi h ep oducible e e ences ha unequi ocally ep esen each e m (Lawless and Heymann, 2010). This classical app oach has h ee main limi a ions. The i s limi a ion is he cos ; DA is a e y expensi e echnique, mainly due o he highly quali ied panel leade s, he de elopmen o e e ences, and he in es men needed o ain he panel in e ms o epea abili y, ep oducibili y and consis ency. The second limi a ion is ela ed o linking DA and hedonic da a which assumes ha he ained panel and consume s ha e simila pe cep ions o he s udied p oduc s (Va ela and A es, 2012). The hi d limi a ion is ela ed o he use o DA o complex p oduc s such as wine (Campo e al., 2010). In DA, di e ences among p oduc s a e iden i ied by measu ing monodimensional a ibu es, while pe cep ion is mul idimensional. This meaning ha pe cep ion is he esul o he in eg a ion o senso y signals elici ed by di e en senso y modali ies o dimensions ( isual, ol ac o y, as e, igeminal) a he han he sum o indi idual signals as hose measu ed by DA (P esco , 2015). This can make in some cases o neglec senso y di e ences in complex p oduc s. In he las decade, al e na i e desc ip i e echniques ha e gained p ominence. Thei main ad an age is o educe aining ime o e en o a oid i because hei pu pose is no o each a consensus among pa icipan s bu o iden i y senso y di e ences among p oduc s, ela ed o ei he monodimensional o mul idimensional pe cep ion (Valen in e al., 2012). Mo eo e , pa icipan s desc ibing he p oduc s a e consume s, which elimina es o educes he di e ences in pe cep ion be ween consume s (hedonic a ings) and ained panellis s (senso y p o ile) inhe en o classical app oaches. The hypo hesis behind hese s a egies is ha consume s can desc ibe p oduc s (Va ela and A es, 2012). In gene al, al e na i e echniques a e mo e apid and, in ce ain occasions, can be good al e na i es o classical DA; bu canno subs i u e DA when an accu a e desc ip ion is needed ( o ex ensi e e iews conce ning al e na i e desc ip i e s a egies see Valen in e al. (2012) and Va ela and A es (2012)). The main d awback o employing un ained panels o desc ibing p oduc s is he di icul y in e p e ing he e ms gene a ed by hem; consume s use wo ds employed in daily con e sa ions, de ined by Lawless and Heymann (2010) as “e e yday language”. To a oid his p oblem, in ce ain domains, echnical expe s a e selec ed o ca y ou he desc ip ion o p oduc s as hei ocabula y is a p io i easie o unde s and and mo e consensual han he ocabula y o consume s. Senso y analysis in wine is a e y good example o his ype o app oach. Wine expe s, including winemake s, p oduce s, oenologis s and wine c i ics ha e de eloped speci ic language and e ms o desc ibing wines and exchanging in o ma ion among hemsel es (B oche and Dubou dieu, 2001). Howe e , he communica ion capaci y o expe s’ desc ip ions was ques ioned some decades ago when Leh e (1975) demons a ed a lack o unde s anding among wine expe s. Wine expe s ha e de eloped a speci ic ocabula y ha includes “scien i ic e ms” and “e e yday language”. Scien i ic o monodimensional e ms a e also employed by ained panels because hey can be unequi ocally ep esen ed by senso y e e ences such as “ke osene” ( ep esen ed by he molecule TDN), “ iole ” (β-ionone) o “honey” (phenyle hyl ace a e) o a oma, “swee ness” (suc ose), “sou ness” ( a a ic acid) o “bi e ness” (quinine sulpha e) o as es, among o he s. “E e yday language” e e s o he indi idual p e ious expe ience o he expe and hus likely d aws on op-down cogni i e p ocesses (Pa e al., 2015). In ela ion o his las ype o language, i is a challenge o unequi ocally in e p e some concep s no mally employed by expe s when desc ibing wines. Some examples o di icul o in e p e wine concep s a e: “mine ali y” (Rod igues e al., 2015), “complexi y” (Pa e al., 2011), “bouque ” (Pica d e al., 2015) OENO One 2021, 1, 69 - 79 71 © 2021 In e na ional Vi icul u e and Enology Socie y - IVES and “g een” (Sáenz-Na ajas e al., 2018) among o he s. Expe s accumula e an impo an numbe o as ing expe iences and iden i y common senso y pa e ns ha a e e balised by using his ype o language. These ambiguous senso y concep s, which a e usually mul idimensional, a e he esul o he e balisa ion o pe cep ion. In his con ex , i is impo an o unde s and and deciphe hese concep s o u he unde s and pe cep ion. The p esen a icle is ocused on unde s anding he concep o g een wines. This concep has been inconsis en ly and aguely desc ibed by di e en wine dic iona ies: “[…] a la ou in wine gene ally associa ed wi h hose o g ass, moss, o ege ables. Also a la ou ound in wines made om unde ipe g apes. A ce ain amoun o g eenness can be cha ac e is ic o , and, he e o e, posi i e in some a ie als like Sau ignon Blanc. Wi h mos ed a ie als, howe e , ob ious g eenness is conside ed a aul […].” (MacNeil, 2015) o “[…] pejo a i e as ing e m o wine made om g apes ha did no each ull ipeness […].” (Robinson and Ha ding, 2015). Simila ly, he scien i ic li e a u e lacks consensus when i comes o he meaning o “g een wine”. The concep is mainly ela ed o a oma e ms such as g een ege ables o ui y/boxwood (G een e al., 2011), capsicum, g ass, oma o lea and aspa agus (Makho kina e al., 2012), allow-like a oma (Moheka e al., 2017), he baceous a oma (Mozzon e al., 2016; Slabizki e al., 2016), o esh cu g ass and “s emmy and s alky”, his las being exempli ied by ozen blackcu an s and blackbe ies (Sal man e al., 2017). In hese a icles, “g een wine” mainly ela es o a concep o med by a di e si y o e ms selec ed e ospec i ely by expe imen e s. This has con ibu ed o he lack o he consensual de ini ion o he e m g een o desc ibing wines. On he o he hand, Gawel e al. (2000) ca ied ou an exhaus i e s udy o iden i y di e en as e and mou h eel ca ego ies o wine desc ip ions esul ing in he mou h eel wheel. On he wheel, “g een” was included in he main ca ego y “as ingency” and he subca ego y “un ipe”. Un ipe was de ined wi h a ibu es om di e en modali ies, including as e and la ou : “[…] a nega i e hedonic g ouping consis ing o an as ingen eel associa ed wi h excessi e acidi y and associa ed g een la ou no es […].” (Gawel e al., 2000). Mo e ecen ly, ollowing a as ing-based s a egy, a g oup o wine expe s om he Somon ano egion, in Spain, sco ed he le el o “g een cha ac e ” o 54 ed wines. They showed ha “g een cha ac e ” was co ela ed o e ms linked o wo senso y modali ies: a oma ( ege al a oma) and igeminal sensa ions (as ingency) and ha i had a nega i e alence. One weakness o he as ing-based app oach is ha he wo k is limi ed o samples ha a e p e iously selec ed by expe imen e s as examples o g een and no g een wines. This can lead o a o ced sco ing o he a ibu e e en i he cha ac e is no p esen o i he e is no pe cep ible di e ence among samples. Thus, e alua o s end o sco e he cha ac e is ics ha di e mos among he wines; which may no necessa ily be he ocus o s udy. A second limi a ion is ela ed o he a igue gene a ed du ing wine as ing, which is mainly linked o he accumula ion o as ingen and bi e sensa ions. To o e come hese d awbacks, Jose-Cou inho e al. (2015) p oposed a cogni i e- based desc ip i e me hodology conside ing expe s’ long- e m memo y wi hou using ac ual wine as ing. They employed his me hodology o pe o m he senso y p o iling o Po uguese wines. This s a egy was success ul in de ining h ee dis inc geog aphical a eas in Po ugal which a e associa ed wi h speci ic wine senso y p o iles. The cogni i e-based desc ip i e me hodology is p o ing o be a p omising ool o conduc la ge scale senso y s udies; i also elimina es he d awbacks o app oaches ha in ol e wine as ing. Thus, he p esen esea ch aims o p o ide eliable in o ma ion abou he g een wine concep by using a non- as ing men al desc ip i e ask, based on expe s’ long- e m memo y. The di icul y in he in e p e a ion o his senso y concep could be a ibu ed o 1) he mul imodali y na u e o he e m and 2) a lack o consensus in i s use by expe s accumula ing di e en senso y expe iences. Thus, in he p esen wo k, we i s hypo hesised ha he e m is mul imodal, being he esul o he e balisa ion o pe cep ion. Besides, he lack o consensus can be he esul o he dis inc lea ning phenomena among expe s; he pe spec i e o he pa icipan s is o en o med in he egion whe e hey wo k as winemake s. This is based on he possibili y ha winemake s om one egion usually a end simila educa ion sou ces and as ing e en s, u he mo e, hey a e mos p obably exposed o simila wines, in con as o expe s om di e en egions. These common expe iences wi hin winemaking egions can lead o he de elopmen o a common e e en ial lexicon o desc ibe and e balise senso y expe iences and concep s (Temme man, 2017). In his con ex , he second hypo hesis o his a icle is he p esence o a egional e ec on expe s’ unde s anding o he g een wine concep . To e alua e his hypo hesis, he concep o © 2021 In e na ional Vi icul u e and Enology Socie y - IVES 72 OENO One 2021, 1, 69 - 79 Ma ía P. Sáenz-Na ajas e al. g een wine has been s udied in ou di e en denomina ions o o igin in Spain by a desc ip i e ask based on long- e m memo y. The e o e, he desc ip ion o a eal wine om memo y (i.e., desc ibe he las g een wine you as ed), which will be di e en o each pa icipan , was employed o gene a e a mo e ealis ic desc ip ion o he ca ego y, han app oaches applied o he iden i ica ion o a cen al endency s o ed in mind, and hus o heo e ical wine p o o ype (e.g., wha is o you a g een wine?). MATERIAL AND METHODS 1. Pa icipan s Se en y-se en Spanish winemake s (65% men and 35% women, aged be ween 23 and 67, a e age o 41 yea s old), om ou wine egions, wi h wo o 43 yea s’ expe ience in hei egion (a e age o 16 yea s) we e in e iewed. The ou egions in No he n Spain mainly p oduce ed wines (> 80% o o e all p oduc ion in all cases). These ou egions we e selec ed because hey p esen ma ked di e ences in he wines p oduced wi hin hem. Thei main di e ences a e hei geog aphical loca ions, clima e, soil p ope ies and a ie y. The DO Somon ano egion p oduces wines wi h a wide ange o a ie ies, including o eign and na ional cul i a s such as Cabe ne Sau ignon, Me lo and Temp anillo; he la e being he mos impo an a ie y. The o he h ee egions a e p oduce s o one main na ional a ie y: Ga nacha in DO Campo de Bo ja and Temp anillo in bo h DOCa Rioja and DO Ribe a del Due o. These las wo egions ha e p ominen di e ences in bo h clima ic condi ions and soil cha ac e is ics. Table 1 illus a es he demog aphic da a o pa icipan s. The panel o pa icipan s did no p esen signi ican di e ences in e ms o gende and age. Winemake s om DO Somon ano had signi ican ly mo e yea s o expe ience (a e age = 20) han hose om DO Ribe a del Due o (a e age = 12). 2. P ocedu e Ques ionnai es we e sen by email o LAAE´s da abase o winemake s. The ques ionnai e included wo main pa s. The i s pa consis ed o a non- as ing ee desc ip ion ask e oking he las g een wine hey had as ed: “Remembe he las ime you as ed a g een wine, could you please desc ibe i ?” In he second pa o he ques ionnai e, pa icipan s we e asked i hey a e used o as e g een wines and i hey belie e ha wines wi h g een cha ac e a e a p oblem in hei egion. Las ly, demog aphic ques ions we e asked (gende , age, wine egion and yea s o expe ience as winemake s). 3. Da a analysis One co pus was ob ained o each egion. Te ms we e lemma ised and synonyms we e eg ouped. Then, he e ms we e g ouped in ca ego ies and subca ego ies o compa e hei equency o ci a ion ac oss egions by using a iangula ion p ocedu e (Ab ic, 2003). Th ee expe imen e s g ouped e ms belonging o simila ca ego ies and subca ego ies. When disc epancies we e ound, hey we e discussed un il consensus was eached. Finally, he equencies o occu ence o each indi idual e m, ca ego y and subca ego y we e calcula ed o each g oup o wine expe s. Only e ms men ioned by a leas 10% o he pa icipan s in a leas one egion we e included in u he analyses. The e ec o egion on he equency o ci a ion o e ms, subca ego ies and ca ego ies was assessed by means o Chi-squa e (χ2). Ma ascuilo pos - hoc pai wise compa isons (95%) we e ca ied ou o signi ican e ec s. The simila i ies and di e ences among he ou denomina ions o o igin we e isualised using a co espondence analysis pe o med on he e m equency able. Analyses we e ca ied ou using XLSTAT ( e sion 19.03, Addinso , New Yo k, USA) and SPAD ( e sion 5.5, CISIA-CESRESTA, Mon euil, F ance) so wa e. TABLE 1. Gende , age anges and a e age age (in b acke s) and ange and a e age (in b acke s) yea s o expe ience in p oducing wine. Signi icance (P- alue) o each a iable (ns: no signi ican di e ences). Di e en le e s in a ow mean signi ican di e ences be ween egions o a gi en a iable acco ding o pos -hoc es s (Fische es ). DOCa Rioja (n = 19) DO Somon ano (n = 19) DO Campo de Bo ja (n = 19) DO Ribe a del Due o (n = 20) P gende (men/women) 9/10 14/5 15/4 12/8 ns age (yea s) 26 – 67 (41) 5 – 30 (44) 3 – 43 (42) 2 – 26 (37) ns expe ience (yea s) 4 – 35 (15)ab 2 – 26 (20)a3 – 43 (18)a5 – 30 (12)b< 0.05 OENO One 2021, 1, 69 - 79 73 © 2021 In e na ional Vi icul u e and Enology Socie y - IVES RESULTS AND DISCUSSION 1. Ca ego ies, subca ego ies and e ms linked o he concep o g een wine Figu e 1 illus a es a hie a chical classi ica ion o e ms employed o desc ibe g een wines. A o al o 23 e ms we e ci ed by ≥ 10% o pa icipan s in a leas one egion (Table 2). These indi idual e ms we e classi ied in wo main ca ego ies: 1) in insic wine cha ac e is ics ela ed o senso y p ope ies (ci ed by 69% o expe s, as shown in Table 2) and 2) o igin o g een wines (ci ed by 60% o expe s). These wo e m ca ego ies, employed o desc ibe he g een wine concep , included six and i e subca ego ies o senso y and o igin, espec i ely. Fo he senso y ca ego y, he ollowing subca ego ies could be iden i ied: a oma, as e, igeminal, colou , mul imodal and hedonic. These subca ego ies co espond o he usual s uc u e ollowed by expe s when desc ibing wine and ecen ly de ined as “wine as ing sc ip ” in an analogy o compu e science sc ip (Hono é-Chedozeau e al., 2020). The ollowing desc ip ion is an example o he wine as ing sc ip : “a wine wi h low colou in ensi y, young, he baceous a omas and in he mou h as eless, as ingen , acidic, in sho , unbalanced. I was an unpleasan wine”. The s uc u e o desc ip ions ca ied ou by expe s s a s wi h a desc ip ion o he colou (low colou in ensi y), ollowed by a oma (he baceous a omas), as e and mou h eel sensa ions ( as eless, as ingen , acidic), hen he use o global mul imodal desc ip o s (unbalanced) and inally an o e all hedonic judgemen o he wine (unpleasan ). By he occu ence o hese wo main ca ego ies, senso y p ope ies and o igin, we see ha expe s ollow a combina ion o bo om-up and op- down cogni i e s a egies o desc ibe wines as was ecen ly obse ed o bo h pe cep ual and concep ual ca ego isa ion o Beaujolais wines (Hono é-Chedozeau e al., 2020) and p e iously epo ed by se e al o he au ho s (Pa e al., 2003; Solomon, 1997). The bo om-up s a egy is based on he su ace cha ac e is ics o he p oduc (i.e., senso y cues), ha is: “ as ing be o e hinking”, while he op-down s a egy e e s o he pa icipan ´s p e ious knowledge and expe ience (i.e., he o igin o he g een wine concep ): “ hinking be o e as ing” (Hono é- Chedozeau e al., 2019). This means ha wine expe s in eg a e bo h cogni i e p ocesses when FIGURE 1. Hie a chical classi ica ion o common ca ego ies, subca ego ies and speci ic e ms ela ed o he g een wine concep e oked by wine expe s. A eas a e p opo ional o he equency o ci a ion o each ca ego y, subca ego y and e m. © 2021 In e na ional Vi icul u e and Enology Socie y - IVES 74 OENO One 2021, 1, 69 - 79 Ma ía P. Sáenz-Na ajas e al. TABLE 2. F equency o ci a ion (exp essed in %) o e ms, ca ego ies (in insic/senso y, o igin) and subca ego ies elici ed in he ou egions de i ed om he desc ip i e ask, o al equency o e ms (%) ci ed by all winemake s (77 pa icipan s), and signi icance (P) calcula ed om chi-squa e es (ns: non- signi ican ; *P < 0.1; **P < 0.05; ***P < 0.001). Signi ican e ms ma ked in bold. Di e en le e s o a gi en a ibu e mean signi ican di e ences based on pai -wise pos -hoc Ma ascuilo es . CATEGORY/ Subca ego y/Te m DOCa Rioja (%) DO Campo de Bo ja (%) DO Somon ano (%) DO Ribe a del Due o (%) TOTAL (%) P 1. INTRINSIC/SENSORY 79 53 74 70 69 ns 1.1. A oma 37 26 42 60 42 ns ege al 26 11 32 25 23 ns he baceous 16 5 11 20 13 ns aspa agus 0 0 0 15 4 ns 1.2. Tas e 53a 16b 42ab 20ab 32 ** excessi e sou ness 32a 5b 26ab 5b 17 ** bi e 26 11 11 15 16 ns 1.3. T igeminal 53a 16b 42ab 30ab 35 * d y 21 5 16 5 12 ns as ingen 26a 5ab 16ab 0b 12 ** low body 11 5 5 15 9 ns ha d/agg essi e 11 011 5 6 ns 1.4. Colou 21 16 21 5 16 ns low colou 11 0 16 0 6 ns ed wine 11 5 5 5 6 ns 1.5. Mul imodal 21 16 32 5 18 ns unbalanced wine 21 16 21 5 16 ns low complexi y 5 5 11 0 5 ns 1.6. Hedonic 32 16 26 10 27 ns unpleasan /de aul 26 11 26 10 18 ns low quali y 0 0 16 0 4 ns 2. ORIGIN 63 42 68 65 60 ns 2.1. Wine componen s 37b 0c 32b 75a 36 *** annins 26a 0b 32a 10ab 17 * py azines 11 0 5 20 9 ns 2.2. Wine ageing 32 32 32 45 35 ns un ipe/young wine 32 32 32 45 35 ns 2.3. G ape a ie y 11 32 16 10 17 ns Cabe ne 5 21 11 10 12 ns 2.4. G ape ma u i y 21 11 16 20 17 ns un ipe g ape 11 511 15 10 ns no phenolic ma u i y 11 11 0 5 6 ns 2.5. Vin age 16 0 11 20 12 ns in age 2015 0 0 11 10 5 ns in age 2018 16 0 0 5 5 ns OENO One 2021, 1, 69 - 79 75 © 2021 In e na ional Vi icul u e and Enology Socie y - IVES e balising he sensa ions p oduced by a p oduc . In he p esen case, he concep o g een wine ac i a es a pu e desc ip ion s a egy o which he pa icipan s use senso y- ela ed e ms as well as a highe -o de cogni i e p ocess in which hey make a diagnosis based on hei echnical expe ise. This op-down p ocess de eloped wi h expe ise can be conside ed an ad an age o e ained panels o consume s when he objec i e o he ask is o p oduce an elabo a ed desc ip ion alongside a diagnosis o he p oblem. Howe e , ained panels o consume s should be p e e ed when a su ace desc ip ion wi hou cogni i e in e ence is he aim, mainly because he panellis s and he consume s end o ollow he bo om- up s a egy (Hono é-Chedozeau e al., 2019; Pa e al., 2003; Solomon, 1997). Fo he g een wine concep , a oma (42%) was he senso y subca ego y ci ed mos , i was ollowed by igeminal (35%), as e (32%) and hedonic (27%) subca ego ies; illus a ed in Table 2. Mo e speci ically, “ ege al a oma” (ci ed by 23% o o al wine expe s), and he hedonic e m “unpleasan /de aul ” (18%) we e among he mos equen ly ci ed ollowed by wo as e- ela ed e ms: “excessi e sou ness” (17%), and “bi e ” (16%). Alongside senso y- ela ed e ms, he o igin ca ego y p esen ed high o e all ci a ions o he wine componen subca ego ies (36%) and wine ageing (35%) wi h hei espec i e e ms “ annin” (17%) and among he mos ci ed, “un ipe/young wine” (35%). No ably, “ annin” was always accompanied by an adjec i e, namely “agg essi e annin”, “g een annin” o “un ipe annin”. The ela i ely low equency o ci a ion o e ms is ela ed o he na u e o he app oach because pa icipan s we e asked o desc ibe he las g een wine hey as ed, and hus o exempli y he ca ego y o g een wines, a he han sea ching o he desc ip ion o a p o o ype, which would ha e led o highe equencies o ci a ions. The mul idimensionali y obse ed o he g een wine concep in he p esen esea ch coincides wi h he esul s obse ed in a s udy ca ied ou by expe s om DO Somon ano which in ol ed he as ing o 54 ed wines (Sáenz-Na ajas e al., 2018). In ha a icle, “g een cha ac e ” o wines was sco ed and signi ican ly co ela ed wi h he a oma e m “ ege al a oma”, he igeminal e m “as ingency” and wi h a e m linked o wine componen s such as (d y and g een) “ annins”. Besides his, “g een cha ac e ” was nega i ely co ela ed wi h p e e ence. On he one hand, hese esul s con i m ou i s hypo hesis ela ed o he mul idimensionali y o he g een wine concep . Thus, i seems o be easonable o hink ha la ou pe cep ion is he esul o he in eg a ion o independen senso y inpu s, mainly a oma, as e and igeminal sensa ions, a he han he pe cep ion o independen s imuli. This in eg a ion leads one o o m a global concep ual uni (P esco , 2015), in his case, he g een wine concep . This in eg a ion o s imuli could be he esul o he mul idimensionali y a ibu ed o a wide ange o senso y concep s such as mine ali y (Pa e al., 2015), complexi y (Pa e al., 2011) o quali y (Sáenz-Na ajas e al., 2016) among o he s. P esco (2015) a ibu ed la ou in eg a ion o epea ed co-exposu e o simila senso y p ope ies o p o iles o a gi en p oduc , and hus o associa i e lea ning. In his con ex , he g een concep could ha e been buil o sha e and communica e expe s´ sensa ions (Temme man, 2017) o conc e e senso y p o iles ha encompass mul idimensional sensa ions co-occu ing epea edly in wines; e balised as g een. F om a me hodological s andpoin , he long- e m memo y s a egy employed in he p esen esea ch yielded e ms simila o hose appea ing in classical app oaches while e oking new e ms. The e ms common o bo h s a egies a e “ ege al a oma”, “as ingency”, “ annin” and he hedonic e m “unpleasan ”. The new e ms we e linked o he senso y ca ego y and as e subca ego y such as “excessi e sou ness” and “bi e ” as well as he o igin ca ego y including he subca ego ies “wine ageing”, “g ape a ie y”, and “ma u i y”. 2. E ec o expe ´s egion on he e ms linked o he g een wine concep Among he 23 e ms gene a ed (Table 2), ou we e he mos ci ed in he ou egions. Among hem h ee belong o he senso y ca ego y: “ ege al a oma” (23% o e all ci a ions), ollowed by “unpleasan /de aul ” (18%), and “bi e ” (16%) whe eas, he ou h, “un ipe/young wine”, belongs o he o igin ca ego y. Besides hese commonali ies, he o he e ms a ied om one egion o he o he one. The signi ican di e ences obse ed o h ee speci ic e ms (Table 2) we e no ewo hy. Two we e senso y- ela ed e ms linked o as e (“excessi e sou ness”) and igeminal subca ego ies (“as ingen ”), and one belonged o he o igin ca ego y and wine componen s subca ego y (“ annin”). “Excessi e sou ness” and “as ingen ” we e mainly ci ed by © 2021 In e na ional Vi icul u e and Enology Socie y - IVES 76 OENO One 2021, 1, 69 - 79 Ma ía P. Sáenz-Na ajas e al. expe s om Rioja and Somon ano, bu ba ely used by pa icipan s om he o he wo egions. The e m “ annin” was ne e ci ed in DO Campo de Bo ja, and highly ci ed in DO Somon ano and DOCa Rioja. The simila i ies and di e ences among egions a e syn hesised in he Co espondence Analysis (CA) plo (Figu e 2). The i s wo ac o s, explaining 82.9% o he o iginal a iance, we e e ained based on Kaise c i e ia (eigen alues highe han a e age). The i s ac o (53% o o iginal a iance) places DO Ribe a del Due o in opposi ion o he o he h ee egions. Expe s om his egion a e cha ac e ised by he usage o he “wine componen s” subca ego ies, which hey ci ed signi ican ly mo e o en (75%, P < 0.001) han expe s om he o he egions, mo e speci ically he e m “py azines” was employed. The second ac o , which explains almos 30% o he o iginal a iance, places DO Campo de Bo ja on he opposi e side o he o he h ee egions. This ac o is mainly explained by he subca ego y “ a ie y” and he speci ic e m “Cabe ne ”, bo h belonging o he ca ego y o igin. This could be a ibu ed o an e ec o highe exposu e and hus he expe s´ amilia i y wi h wines elabo a ed wi h his a ie y om his egion. Cabe ne is mainly cul i a ed in DO Somon ano ollowed by DO Campo de Bo ja. The lowe ci a ion o he e m Cabe ne in DO Somon ano may be because hey cul i a e a wide ange o a ie ies ha a e associa ed wi h he g een wine concep such as Mo is el o Me lo in compa ison o DO Campo de Bo ja whe e Cabe ne was he sole a ie y ci ed. Desc ip ions o winemake s om DOCa Rioja and DO Somon ano seem o be mo e con e gen , wi h a a he high emphasis on as e (“excessi e sou ness”) and igeminal (“d y” and “as ingen ”) senso y subca ego ies. DO Campo de Bo ja p esen s an in e media e posi ion wi h a close p ojec ion o DOCa Rioja and DO Somon ano. This is mainly due o he ela i e impo ance gi en o common ca ego ies ela ed o he o igin o g een wine, such as he “wine ageing” subca ego y (“un ipe/young wine”) and senso y- ela ed e ms linked o “colou ” and he mul imodal e m “unbalanced wine”. These esul s demons a e ou second hypo hesis, ha e en hough he e a e common e ms, sha ed by expe s om he ou egions, which belong o he senso y ca ego y and a oma, as e and hedonic subca ego ies (“ ege al a oma”, “bi e ” and “unpleasan ”) an impo an idiosync asy is linked o he g een wine concep ha is con ingen on he egion o o igin o he wine expe s. Ou second hypo hesis is demons a ed by he ac ha he egion a ec s he de elopmen o senso y concep s by wine expe s as he one s udied in he p esen esea ch (g een wine). This esul could be a ibu ed o he expe s´ dis inc exposu es o g een wines and hus o di e en le els o FIGURE 2. Co espondence Analysis space p ojec ion o ca ego ies, subca ego ies and e ms elici ed by pa icipan s (ci ed by a leas ≥ 10% o pa icipan s in a leas one egion) om he ou egions on Fac o s 1–2. OENO One 2021, 1, 69 - 79 77 © 2021 In e na ional Vi icul u e and Enology Socie y - IVES amilia i y wi h his concep . Howe e , his does no seem o be he case because he sel - epo ed incidence o g een wines p esen ed no signi ican di e ence among egions (45 - 60% o pa icipan s acknowledged he p esence o g een wines in hei egions). Al e na i ely, di e ences in he concep o g een wine among egions could be ela ed o a ia ions based on dis inc echnical and senso y expe iences acqui ed by expe s om di e en egions. Zannoni (1997) s a ed ha he same wo d can be associa ed wi h di e en concep s depending on he loca ion because i is linked o he cul u e o indi iduals. He cle e ly showed he example o he wo d “b ead” ha can be easily ansla ed in o di e en languages (“pan” in Spanish, “pain” in F ench, “B o ” in Ge man, o “pane” in I alian). Howe e , he concep associa ed wi h b ead is o ally di e en (in e ms o ex u e, o m, composi ion…) o indi iduals belonging o di e en cul u es. Thus, expe s belonging o di e en p oduc ion egions accumula e di e en expe iences which can in luence he o ma ion o senso y concep s. Thus, one can assume ha op-down cogni i e p ocesses a e used o o m knowledge om expe ience, which would gene a e di e ences among expe s om di e en egions. Along hese lines, Pa e al. (2015) epo ed ha mul idimensional concep s (complexi y, amilia i y o liking), as could be he case o g een wine, show g ea e di e ences in hei meaning among di e se cul u es (unde s ood as social g oups exposed o di e en expe iences), han simple analy ical desc ip o s which a e employed o e balise speci ic a oma o as e concep s. This di e ence is ela ed o he dis inc p ocesses in ol ed in he o ma ion o bo h ypes o concep s. Mul idimensional concep s a e mainly ela ed o op-down s a egies, while analy ical desc ip o s co espond mo e o bo om- up p ocesses. CONCLUSIONS The p esen wo k has shed ligh on he concep o g een wine which is concep ualised wi h a long- e m memo y s a egy. To s a wi h, he esul s con i m ha expe s use a combina ion o bo om- up and op-down s a egies o ca y ou wine desc ip ions; hei desc ip ions a e suppo ed by he use o bo h senso y and o igin ca ego ies o desc ibe wines. Secondly, he mul idimensionali y o he g een wine concep has become e iden in ha i includes a oma, as e, igeminal, colou , mul imodal and hedonic senso y subca ego ies. Thi dly, he senso y dimensions “ ege al a oma”, “bi e ” and “unpleasan ” a e he speci ic e ms sha ed by expe s om he ou egions o desc ibe g een wines and may ep esen he co e o he g een wine concep . Despi e hese commonali ies, an impo an egional e ec ela ed o senso y e ms associa ed wi h as e (“excessi e sou ness”) and igeminal (“as ingen ”) sensa ions and wine componen s (“ annin”) was demons a ed. This e ec is a ibu ed o he op-down p ocess in ol ed in he concep ualisa ion o g een wine. Thus, echnical and senso y expe iences di e ing among expe s om dis inc wine egions can con ibu e o he o ma ion o di e se wine concep s. Thus, mo e s udies e alua ing he g een wine concep in o he wine egions a e equi ed o unde s and he senso y concep and imp o e communica ion among wine audience. The esul s illus a ed ha his app oach can deciphe and p omo e he cha ac e isa ion o ill-de ined concep s used by winemake s. The main ad an ages o memo y-based app oaches o e classical me hods a e ha hey do no equi e he in as uc u e needed in classical as ing app oaches, and hey consume a less ime and esou ces. A po en ial limi a ion o he p esen app oach is ha pa icipan s we e asked o emembe he las ed wine hey as ed, which leads o low equencies o ci a ions o speci ic e ms. The idea behind his app oach is ha he cha ac e isa ion o di e en exempla s s o ed in he memo y o expe s will p o ide a be e desc ip ion o he ca ego y, being close o eali y, han he desc ip ion o a cen al endency o a p o o ype o wine, which p o ide a mo e heo e ical cha ac e isa ion, bu leads o highe equencies o ci a ions. Acknowledgemen s: This esea ch p ojec was unded by he Spanish Minis y o Economy and Compe i i eness (MINECO) (p ojec AGL2017- 87373-C1/C3). I.A. and S.F.T. a e g a e ul o he MINECO o he p edoc o al ellowships FPU-2015, and FPI-2018 (Uni e sidad de La Rioja), espec i ely. MPSN acknowledges he Spanish Na ional Resea ch Agency, he Minis y o Science, Inno a ion, and Uni e si ies and he Eu opean Social Fund o he pos doc o al ellowship: Ramón y Cajal P og am (RYC2019- 027995-I/AEI/10.13039/501100011033). LAAE acknowledges he con inuous suppo o Dipu ación Gene al de A agón (T53) and Eu opean Social Fund. The au ho s also exp ess hei g a i ude o he wine expe s o pa icipa ing in he s udy.