Wine Terms

In my last post I mentioned I was using several websites (and pages within those sites) that had English translations to extract side-by-side human English translation of the (presumably) original Spanish. OK, done – so what? Like I’ll be doing with all sources then I begin an extraction process to add pairs (words or phrases) of translations to my corpus. A key part of that also has to be asserted some measure of “certainty” whether the translation is correct. Using a probability type measure (0.0…1.0 obviously fits). Then the corpus analysis program can find as many of the same pair as it can and evaluate a new certainty, i.e. something like – lots of pair instances that are the same but possibly each low certainty may be as good as few of a pair with high certainty. An interesting question, then, is human translation (relatively rare) of websites (mostly menus) more reliable source of information than machine translation.

Of course the extraction process itself (which I do and therefore is subject to error) plays a role as well so I’ll use my small corpus of wine webpages to extract a set of pairs and then use any other sources of wine terminology to confirm/deny my pairs (just manually, so I understand the data, before trying to write code to do this). So here’s my result:    (scroll down past list for more of this post)

Abierto 2 open
Acerb 2 acerbic
acidez 1 acidity
ácido 1 acid
Aciete Esencial 2 essential oils
afinamiento 1 refinement
afrutado[s] 1 fruity
agradables 1 nice, pleasant, agreeable
Alegre 2 zingy
Amoratado 2 inky
Amplio 2 big
Añada 2 vintage year
arcillo 1 clay
Armónico 2 harmonious
aromas 1 aromas
aromática 1 aromatic
Barrica Bordelesa 2 Bordeaux cask
barrica 1 cask or barrel
beber 1 to drink
Blanco Seco 3 Dry White
Blanco 2 white
boca 1 literally mouth, but can mean palette in wine tasting context
Bodega 3 winery
Bodeguero 3 winemaker
Bota 2 butt
botella 1 bottle
Botritis 2 botrytis
brillante 1 bright
brotaciones, brotación 1 [not found] budding ? (derivative of brotar)
brotar 1 to sprout, bud
calidad 1 quality
campaña 1 growing period, season
campo 1 field
canela 1 cinnamon
cánones del clasicismo riojano 1 classic Rioja style (not literal)
Capa 2 layer
cata 1 tasting (action of)
cereza 1 cherry
Cerrado 2 closed
Clarificación 2 fining
clásico de Rioja 1 Rioja classic
comarca 1 region, district
complejidad 1 complexity
Complejo 2 complex
Corcho Cork
cosecha 1 harvest, crop; vintage
Crianza en barrica 4 Aging in barrel
crianza en madera 1 aged in wood (literally, cask colloquially)
crianza 1 aging
cuerpo 1 body
Dejo 2 aftertaste
Denso 2 dense
Depositos 4 Deposits
Dorado 2 golden
Dulce 2 sweet
Elaborado por 3 Produced, matured by.
elegante 1 elegant
Embotellado por 3 Bottled by
Embotellar 4 To bottle
en barrica 1 in cask or barrel
envejecimiento 1 aging (also laying down)
equilibrado 1 balanced
equilibrio 1 balance
Especiado 2 spicy
Espeso 2 thick
Estructura 2 structure
Evolucionado 2 evolved
expresivo 1 expressive
Fermentación alcohólico 4 Alcoholic fermentation
Fermentación maloláctica 4 Malolactic fermentation
fermentación 1 fermentation
final de boca 1 “finish” (literally end/finish of mouth)
final 1 after-taste
fino 1 fine
florals 1 floral
fresco 1 fresh
frescura 1 freshness
frutos cítricos 1 citrus fruits
Fuerte 2 strong
Graciano 1 red grape variety
grados 1 grade or degree (but alcohol by volume)
Heces 2 sediment
Hoja 4 Leaf
Hollejo 2 grape skin
Joven 2 young (little or no aging)
Jurado de Cata 2 wine tasting panel
Lágrimas 2 tears
Levaduras 4 Yeast
Lías 2 lees
limpio 1 clean
Maceración Carbónica 2 carbonic maceration
Maceración en frío 2 cold maceration
maceración 1 maceration
madera 1 wood
madura 1 ripe, mature
madurar 1 to mature
Manchado 2 literally ‘stained’
manzana 1 apple
maridaje 1 literally marriage or combination; food matches/pairings
Mazuelo 1 red grape variety
mezcla 1 mixture, blend
mosto 1 must (grape juice)
nariz 1 nose (also aroma)
notas 1 notes
olores 1 smell (scents in corpus)
Oro 2 gold
Oxidación 2 oxidation
parámetros de calidad 1 quality indicators
Pasa 2 raisin
Pepita 4 Seed
Perfumado 2 perfumed
Persistencia 2 persistence
Pimienta 2 black pepper
postgusto (posgusto) 1 [not found] after-taste
Prensa 4 Press
prensado 1 pressing
pulidos 1 polished
Rama 2 branch
Recio 2 gutsy
Redondo 2 rounded
Refrescar 2 refresh
Regaliz 2 liquorice
Roble Americano 4 American oak
Roble Francés 4 French oak
roble 1 oak (as in the barrels)
Rojo 2 red
Rosado 2 rosé
sabor 1 flavor, taste
Sabroso 2 flavorsome
Seco 2 dry
sedoso 1 silky
Semidulce 2 semi-sweet
Semiseco 2 semi-dry
sensación 1 sensation
Suave 2 smooth
suelos 1 soils (also ground, floor, land)
Tabaco 2 tobacco
tanino 1 tannin
temperatura controlada 1 controlled temperature
temperature de servicio 1 serving temperature, aka, best served at
Tempranillo 1 grape variety
terciopelo 1 velvet
Típico 2 typical
trasiegas 1 decant (rackings in corpus)
untuoso 1 literally greasy (aka unctuous), but nicer means ‘smooth’
uva 1 grape
Vainilla 2 vanilla
valores 1 values (as in levels of an indicator)
variedad 1 variety or varietal
vendimia 1 vintage, grape harves (whole process)t
Vid 4 Vine
Vina 3 Vineyard.
viñedos 1 vineyard, vines
Vino blanco 4 White wine
Vino de calidad (Quality wine) 3 Must come from a DO or DE. Only wine made from the free-run or lightly pressed juice of ripe healthy grapes, which has undergone a temperature controlled fermentation, qualifies.
Vino de cosecha, or vendimia 3 Wines of a particular vintage year. In special cases, if the purpose is to improve the quality of the wine, a maximum of 15% of wine of a previous year may be added.
Vino espumoso 4 Sparkling wine
Vino Fino de Mesa 3 Fine table wine.
Vino Generoso 3 Special aged dry or sweet wines of higher alcoholic strength than table wines. From the Latin term for excellence. Sherries are vinos generosos.
Vino rosado 4 Rosé wine
Vino tinto 4 Red wine
vino 1 wine
Viura 1 white grape variety
viveza 1 vividness, strength
Vivo 2 lively
Yema 2 yolk
Zarzamora 2 blackberry

I combined four lists. In MSWord I can use different colors and fonts for each list so when I merge them I can easily see where any pair came from, but here in WordPress formatting is more limited so the middle column indicates the source. My extracted list (from all those webpages I processed from both bodegas and restaurants) is 1.  I choose not to provide links for the other three sources, but 2 was certainly the largest.

I eliminated duplication and then used a simple notion of “certainty”. Items from list 1 that are shown here in bold had one or more identical (or almost identical) translation in one of the other lists. This isn’t particularly robust definition of certainty but it will do for this proof concept.

So of the 171 terms in the merged list (82 are from my manual extraction, the remainder from one of the other three lists) only 24 of my extracted terms get marked as “certain” due to occurring in other lists:

afrutado[s], barrica, botella, cata, cosecha, elegante, equilibrado, fermentación, final de boca, fresco, maceración, madura, mosto, postgusto (posgusto), roble, sabor, sedoso, tanino, untuoso, uva, variedad, vendimia, viñedos, vino

There could have been some more since I did not extract really obvious terms from my corpus, such as blanco or seco or dulce or uva. And two of the “confirmed” terms actually are in dispute. Once source admits afrutado is used for ‘fruity’ but this is actually wrong and the term should be frutal. The dictionary confirms afrutado does mean ‘fruity’ but this does not confirm it is the correct term to use in a wine context. Likewise it confirms frutal to be fruit or fruit tree but doesn’t mention how this would be a taste term for wine. So who knows? Which is right? Wine terminology (in English) sometimes contradicts the more common meanings of words since wine tasters understand a particular word in a particular context (and we amateurs just have to learn what they mean). So it’s certainly possible this source might be right BUT how would this ever be confirmed.

Likewise postgusto (clearly ‘after taste’ from context) doesn’t appear in any dictionary. And, in the other lists it appears but is spelled posgusto. Now I’m not sure if this meets the definition of neologism, especially as ‘post’ can mean ‘after’ (in this context) in English but doesn’t occur in Spanish whereas is ‘taste’ or ‘flavor’ so does this word actually exist (or get used in wine documents) and which is the appropriate form?

There was also some conflict between viñedos and vina.  Both are in the dictionary as vineyard but only vina is listed as vines. That is then potentially a flaw in my extraction of pairs since I saw viñedos clearly translated as ‘vines’ in a human translation, but, of course, that person may confuse these two terms.

The term I’m happy I was able to figure out (lots of examination of text to reach my conclusion) is final de boca. This literally would translate to ‘end of mouth’. but it’s more accurate to translate it as ‘finish’, which is actually one of those terms where its usage in wine descriptions has quite different meaning than its common meaning. And one of the lists pronounced that just final is sufficient for ‘finish’ which is one of the literal translations itself. OTOH boca itself has some ambiguity.  It literally means ‘mouth’ but was commonly translated as ‘palette’ in the human translations. That’s not any of the literal translations of ‘palette’. But, again, palette is a word that has different meaning in wine tasting context than its more common meanings.

So, this is all human analysis, with a lot of trial-and-error, back-and-forth, looking in dictionaries and doing web searches. In this contest of John Henry and the machine I think man will win so I really wonder how effective any AI (or just statistical analysis) can be. OTOH, ‘man’ needs to be a fluent Spanish speaker who participates in Jurado de Cata (wine judging panel) and I fall way short of that. But, still, what is the chance I can still produce the best list of wine terms freely available on the Internet? Pretty good, I’d say (given few are even trying).



Something different – wine label and description

By coincidence I decided to get some good wine for our Valentine’s Day dinner. We cook ourselves because: a) we actually can cook some things better than restaurants do, and, b) we spend our money on the ingredients, not the restaurant’s labor and real estate. So off to Whole Foods for some very good wines at the same price as medium wine with restaurant markup. There isn’t a lot of Spanish wine available here. Trader Joe’s has some amazing values, cheap but tasty Spanish wines, but for a Reserva Whole Foods was my only option. Since I bought this wine I’ll allow myself to link their image.




from Bodegas de los Herederos del Marqués de Riscal (website)



Visit the website link I provided as this is an interesting place, very oriented to visitors and with a striking Frank Gehry designed hotel and elegant restaurant.

But finding this wine, first in my little used PeñínGuide to Spanish Wine 2016 which led to the website, gave me an opportunity to look at some translation issues related to wine. For the wine I bought there is a PDF for Spanish and another for English which appears to me to definitely be a human translation, thus providing the rare opportunity to compare side-by-side Spanish, human English, and computer English. For example:

Antes de salir (lit: go out, leave) al mercado tiene (lit: has) un period mínimo de afinamiento (lit: refinement) en botella de un año. Before release for sale it spends a minimum of one year rounding off in the bottle; time enough to show how much complexity tempranillo is able to achieve.

{Before going on the market, it has a minimum bottle-tuning period of one year.}

[Before going on the market, it has a minimum refining period in bottle of one year.]

I did a few dictionary lookups and noted the translation in the Spanish as (lit: whatever). The first English translation is the human one directly from the PDF. This has a definite clue that it’s human translation since the English includes an additional part (underlined) that has no match of any kind in the Spanish so the author chose to add this bit.  The {whatever} part is the translation done by (actually Microsoft) and the [whatever] part is the translation done by Google (had to paste the Spanish in my own test page at this blog since PDF’s don’t get processed by Google in Chrome).

For me there are a couple of interesting issues in these translations:

  1. ‘Before going on the market’ seems to be a more “accurate” translation of Antes de salir al mercado BUT the human translation “Before release for sale” might actually be more accurate, i.e. this wine might not have literally gone to a mercado in order to be sold.
  2. period mínimo de afinamiento en botella is interesting to see the three different corresponding English: [human] “minimum of one year rounding off in the bottle”, [spanishdict] “minimum bottle-tuning period”, and [Google] “minimum refining period in bottle”. When I look up afinamiento I get refinement which Google uses (also the closest to word-by-word literal translation); I think this is definitely better than ‘tuning’ (no idea where that came from) and perhaps better than the human ’rounding off’ ‘period’ is omitted in the human translation but literally present in Spanish and both machine translations.

So let’s look at some more for this, some simple differences in the human translation versus literal lookup or machine translation:

VARIEDAD DE LA UVA (lit: variety of grape) VARIETY USED
GRADOS (lit: degree or grade) 14,1º ALC./VOL 14,1º

Grados is probably not a translation issue, just a different description used in Spain versus the more typical one used in U.S. (although note the English is British, not U.S. English so who knows what this might mean, as in possibly a legal labeling requirement somewhere).

MARIDAJE (lit: marriage, combination,  union) FOOD MATCHES

And this is another interesting turn of phrase. In U.S. “food matches” might also be “food pairings” and, in a stretch, “married” might be used in this context. With only this single sample I can’t draw any conclusion but I find it amusing language to use maridaje for this meaning.


Again, the human translation is definitely not very literal but carries the meaning just fine and frankly I’d prefer the English term (which literally translates to the bulky mejor servido en),

ATRIBUTOS (lit: attributes) GUSTATIVOS (lit: taste) APPEARANCE (lit: aspecto o apariencias)

This one, however, is a little misleading (I think) to switch from ‘taste attributes’ to ‘appearance’. The text (see some below) under this heading covers: color, nose, tannin and finish, a mixture of sight, smell and taste sensations so ‘appearance’ is a bit too narrow to cover all these.

En boca (lit: mouth) es fresco, con taninos pulidos (lit: polish) muy agradables (lit: nice, pleasant, agreeable), con buena estructura pero fácil de beber. Fresh and easy to drink on the palate, good backbone and lovely, polished tannins.

{In the mouth it is fresh, with very nice polished tannins, with good structure but easy to drink.}  

[The palate is fresh, with very nice polished tannins, with good structure but easy to drink.]

The human translation, though useful and pleasing, has little resemblance to the original Spanish (backbone is completely missing in the Spanish). The spanishdict translation is quite literal but definitely gets the meaning across (in wine tasting tannin is almost something you feel on your tongue (pucker) rather than a taste). How Google decided to use palate for boca is surprising – perhaps part of their claim their AI figures out translation via context and while dictionary lookups certainly do not have palette for boca or boca for palette it is appropriate and surprising that the machine translation went down the same path as the human translation.

While there are many more interesting things I’m finding from this description webpage I should wind down and so I’ll just leave you with these bits of the description of the weather at the vineyards for this vintage year (spanishdict translations {xxx} added to human translation.

La vendimia de este año ha estado condicionada, en gran medida, por varios puntos clave sucedidos a lo largo de toda la campaña.

Comenzamos el ciclo con un estado de reservas importante, que se tradujo en brotaciones buenas y viñedos con una carga en general elevada.

La ausencia de heladas primaverales, vientos fuertes en brotación y granizadas de verano, hacen que lleguemos a mediados de septiembre con unas uvas muy sanas y con unos parámetros de calidad que sugerían estar ante una cosecha interesante.

This year’s vintage has been, to a great extent, conditioned by a series of key events during the growing period.  {This year’s harvest has been largely conditioned by several key points that have occurred throughout the campaign.}

We started the cycle with good reserves and this was reflected in good budding and vines which would be heavily laden in general.  {We started the cycle with a major reserve state, which resulted in good sprouts and vineyards with a high overall load.}

The absence of spring frosts, strong winds during budding and hailstorms in the summer meant that we reached the middle of September with very healthy grapes and quality indicators which promised a very interesting harvest was on the way. {The absence of spring frosts, strong winds in sprouting and hailstorms of summer, make that we arrive in mid-September with very healthy grapes and quality parameters that suggested to be before an interesting harvest.}

I will crunch this some more (plus extract even more from this website) to obtain a list of useful terms in describing wine.

That, and drool a bit, at the prospect of actually visiting this place and staying at their hotel and chowing down on their menu but short of winning the lottery that probably isn’t going to happen.

P.S. I found a restaurant (website) that carries the wine (above) and so found a price, 23€, which is about $29 and about what I paid at Whole Foods. But that is a restaurant price (with service) so I’d guess a bottle of this wine in retail outlet (or at the winery itself, quite a touristy place) for around $20 or somewhat less than retail imported into U.S.