2

I've found this Worldwide map or data for linguistic distance here when looking for a way to know if Portuguese is the most similar language to Spanish. Unfortunately neither of these languages are in the marked answer. Is it know through this mean or another, which language is the most similar to Spanish?

Pablo
  • 446
  • 4
  • 16
  • 6
    I'm afraid this will break down to a definition of "language". Does South American Spanish count? What about Aragonese or Asturian? Does Valencian count? – Sir Cornflakes Aug 30 '19 at 15:55
  • 1
    As pointed out in the previous question, where you found the map, the data in this map is suspect and certainly not definitive. And as @jknappen points out, "closest" is a matter of opinion and definition, neither of which is certain. – jlawler Aug 30 '19 at 17:25
  • 1
    Besides the great points others make, I would point out that distance ie intelligibility is not necessarily symmetric. Doublets ie loans from related languages or classical forms, conservative spelling, prescriptivism, exposure and so on make speakers of some variants of some languages much more attuned to related languages. The Iberian situation has plenty of such asymmetry. – Adam Bittlingmayer Aug 31 '19 at 14:06
  • And continuing @jknappen's point, if Valencian does count, is it the Valencian of the Acadèmia de la Llengua Valenciana or the Valencian of Lo Rat Penat? – Peter Taylor Sep 03 '19 at 07:47

2 Answers2

9

The map you have doesn't pass the sniff test for me. I don't imagine anyone realistically saying Catalonian being closer to Spanish than Galician. I can't speak for other Romance groups, but for Iberian languages, the chart on Wikipedia is in line with my experience (I speak Asturian, Castilian, and Portuguese, and regularly consume media in Galician and Aragonese).

The problem, however, when doing a lexical analysis is how you handle words that exist in both languages but have different frequencies of use, or with pronunciations that line up closer in one than another.

For example, cerrar (to close) exists in Portuguese, but fechar is used infinitely more. On the other hand, in both Asturian and Aragonese, the verb is zarrar. Which is closer? You can argue for Portuguese because cerrar is identical — but should the comparison really be with fechar which is more used and thus you could argue for Asturian or Aragonese with zarrar?

Or looking at Spanish’s despertar to Asturian's espertar, Portuguese and Catalan's despertar and Mirandese’s spertar which one is closer? Do we count them all the same since it's a direct etymological link and the only difference is the morphological form of *des ? How do we factor in Asturian's esconsoñar and Portuguese and Mirandese’s acordar? Do we link Asturian more closely because the two are about equally common in it but (de)spertar is rare in PT/MWL?

Do we let pronounciation factor in because some languages like Mirandese indicate vowels that would be spoken (but not written) in others like Portuguese? If you do, how do you decide which spoken variant to use? Gato in Portuguese/Mirandese is gato which might make them look closer to Spanish, but oftentimes the pronunciation will trend closer towards /u/ (Mirandese) or even silent (Portuguese) in which case … are they really closer to Spanish than Asturian’s gatu?

If you don't establish a clear methodology for how you're comparing things, then the numbers are meaningless. Most likely I would say that Ladino would be the language closest, but it has a modern writing system that probably obscures that closeness and I'm not familiar enough with it to say definitely. Otherwise, any of the West Iberian languages would be good candidates and I'm not sure you can say confidentally which one is lexically closer than the others for the reasons above.

user0721090601
  • 944
  • 5
  • 12
  • Can you give me your personal feeling about it here in the commentaries? If you speak those languages, and I'm pretty sure you have a good idea of catalan since it's from the area of the languages you know, which one sounds more closer to spanish for you? – Pablo Aug 30 '19 at 19:20
  • 3
    @Pablo not all of the similarity is purely lexical though. I don't believe Catalonian to be the closest by any stretch of the imagination. From a pure "if I listen to it do I understand it" it absolutely would need to be Ladino: https://www.youtube.com/watch?v=6q323m0RRPU . You can compare it to Aragonese ( https://www.youtube.com/watch?v=SkBIv78gEMk ) or to Asturian ( https://www.youtube.com/watch?v=AkSzEtbX9Wo ), Eonaviego/Fala ( https://www.youtube.com/watch?v=TGIVhHJ6e8U ), Mirandese ( https://www.youtube.com/watch?v=-gnJtZFyzZA ) or Galician https://www.youtube.com/watch?v=WDgxTUwEXes – user0721090601 Aug 30 '19 at 20:04
  • 1
    -1 : it's a bit harsh because I like your answer, but your "sniff test" doesn't focus on the way the lexical distances have been computed. You may critic the shortness of the lexic list used for computing the distances, or other things in the method. But your answer carefully avoid to enter in such details. If we look at the original answer mentionning this map, there are some clear details: – Stephane Rolland Sep 01 '19 at 06:41
  • I quote: But I think I have a pretty good idea of how the data was obtained. These maps basically show the Levenshtein distances lexical distance or something similar for a list of common words. Now this list could be the Swadesh № 100 or № 207 list with counting duplicate letter shifts in different words as one LD, or it could be Dolgopolsky № 15 list or a Swadesh–Yakhontov № 35 list and just brutally counting Levenshtein LDs on those lists. Or Tyschenko's could have his own list of words and methods to calculate the lexical distance. – Stephane Rolland Sep 01 '19 at 06:41
  • I have looked at the Swadesh list, and though it is very short, 100 words, it is based on very very very common words. And find this idea of lexic distance valid, though it may not be appropriate for you. – Stephane Rolland Sep 01 '19 at 06:44
  • I mean @Alternative Transport , the original answerer, achieves similar results of Prof. Konstantin Tishchenko . That is a reproducible method. – Stephane Rolland Sep 01 '19 at 06:47
  • In a nutshell I deeply disagree on your conclucion: "If you don't establish a clear methodology for how you're comparing things, then the numbers are meaningless." Because the method is explained. – Stephane Rolland Sep 01 '19 at 06:56
  • What the heck is "Catalonian", some kind of muskmelon?:) Surely the word for the language matches the gentilic, so simply Catalan. – tchrist Sep 01 '19 at 13:07
  • 1
    @StephaneRolland the point is, why limit to a Swadesh list? Do you count Portuguese ça as equivalent to Spanish za or Asturiano dixo as equivalent to Spanish dijo (because, prior to orthographic reform, Spanish used ça and dixo). If not, then you are letting the artifices of writing influence the LD. And of course the words chosen for analysis effect things greatly. So you have two very arbitrary numbers being plugged into an equation. – user0721090601 Sep 01 '19 at 17:15
  • You guessed at the methodology but I'd need a detailed analysis for how it accounts for that. And if it uses Swadesh list, which ones? How do the determine whether Portuguese uses negro/preto or Asturian negru/prietu or Spanish negro/prieto? Those are binary choices that not insignificantly affect results and absent an explanation of how they are accounted for (in addition to the aforementioned orthographic and phonetic issues), I'd not trust the results. – user0721090601 Sep 01 '19 at 17:23
  • The "sniff test" — it seems you may be unfamiliar with the term — refers to whether results even make sense with the known/experienced reality. If someone comes up with a method that says that Basque is closer to Spanish than French, regardless the methodology, the results don't pass the sniff test — anyone with a passing knowledge of the languages involved know the results are simply wrong, and there must be an error in the methodology. I have more trouble reading Catalonian texts precisely because of lexical differences. I'd need to see very solid evidence to say it's closer. – user0721090601 Sep 01 '19 at 17:36
  • @guifa For the methodology used, I am sorry I can't read Russian/Ukrainian: This is the book by Professor Tyshchenko of Taras Shevchenko Kyiv National University where the map is taken from (as far as I have understood) http://linguist.univ.kiev.ua/museum/book/index.html – Stephane Rolland Sep 01 '19 at 18:37
  • @guifa I agree that this metric may not satisfy your definition of being close for languages. I totally disagree with your disdainful judgement that this metric is meaningless. It's a metric, and it has shown some coherent results for distinguishing Languages Families as far as I have read. So NO it does not seem MEANINGLESS. – Stephane Rolland Sep 01 '19 at 18:45
  • @guifa I feel really ill at ease that you use such a FALSE argument implying that the methodology says that Euskera is closer to Spanish than French. The computed lexic distance between Euskera and Spanish is 70, whereas it is 41 for Spanish-French. I stop this conversation unless/until you read the related answers and the related work carefully. – Stephane Rolland Sep 01 '19 at 19:13
  • @StephaneRolland you are intentionally misrepresentating what I have stated in the comments. I'm not going to respond further. Downvote as you wish. – user0721090601 Sep 01 '19 at 19:19
  • @guifa No there is no misrepresentation from me. Don't gaslight. The map says: Eus-Spa 70 / Spa- Fra 41 , you misrepresent the methodoly in saying: "If someone comes up with a method that says that Basque is closer to Spanish than French, regardless the methodology, the results don't pass the sniff test" – Stephane Rolland Sep 01 '19 at 19:27
  • 2
    @StephaneRolland, you seem to be confusing a hypothetical used to illustrate a definition with a claim of fact. guifa didn't say that anyone (much less Tyshchenko) is claiming that the distance Basque-Spanish is smaller than the distance Spanish-French: they said that (1) if someone were to claim that then it would cast doubt on the validity of their method; (2) the claim which Tyshchenko does make that the distance Spanish-Catalan is less than Spanish-Gallego is not as extreme as that example, but still enough to make guifa mistrustful. – Peter Taylor Sep 03 '19 at 07:33
  • @PeterTaylor I understand your point of view. Let me rephrase my understanding shortly: The answer refuting closeness of Catalan compared to Portuguese (based on the lexical distances map) is based on saying that the metric used is MEANINGLESS, because it does not pass the SNIFF TEST. Answerer gives an example of NOT PASSING THE SNIFF test. But this example is indeed covered CORRECTLY by the methodology. Moreover, the Language Families that emerges from this methodology do make sense. My conclusion: the metric is MEANINGFULL, and we should not use outright disdain for concluding things – Stephane Rolland Sep 03 '19 at 09:25
  • @PeterTaylor BUT, I totally understand the answerer point of view that the lexic used for computing the lexical distance may be too short, that there should be other things to take into account to estimate how languages are close. A more complex, more accurate metric is needed to be used. I was thinking about weighing the words with their frequency usage. This is more or less achieved by a short list taking only the 100 most frequent words list. But obviously it needs improvement since speakers of Catalan, Galician, and Portugese feel the metrics seems incorrect/not sufficent in their case. – Stephane Rolland Sep 03 '19 at 09:26
0

NB: The problem with the other answer by guifa is: it claims that there is no methodology behind the lexical distances used in the map in question, and that those numbers are meaningless. I'm ill at ease: Reading entirely the two original answers and their links shows otherwise (though unfortunately I can't read Russian to comment Konstantin Tishchenko's work).

In the map in the marked answer: Catalan language is displayed as the closest to Spanish.

Galician and Portuguese, both are more distant than Catalan.

Looking the original map in Russian by Prof. Konstantin Tishchenko in the other answer, Galician AND Occitan/Provencal are closer to spanish than Portugese.

About the validity and the methods used for the two maps: the answers mentions Levenstein distance, as a systematic method for computing the distance between two sentences. Check those answers for more detailled informations.

The answers suggest that the author Konstantin Tishchenko used only a limited set of very usual words to calculate the Levenstein distance, or a similar metric.

e.g. if you look at the Swadesh list that could have been used, then you notice it is really short a list! Only a hundred of words.

That may be insufficient for saying that two languages are close. Maybe. But it is another matter to state that those distances are meaningless, and I don't subscribe to that point of view.

Stephane Rolland
  • 652
  • 6
  • 18
  • 1
    This is very strange. Even though I can understand something of catalan by speaking spanish, I can't understand as much of it as portuguese. Could there be other factors to consider that those maps aren't counting? – Pablo Aug 30 '19 at 19:14
  • 1
    @Pablo Catalan and Occitan belong to a completely different branch of Western Romance than the one on which one can find Castilian and Asturian, Portuguese and Galician. The Languedoc branch differs from the latter branch by about 20% measured in shared cognates in the overall lexis. We're talking low 70s versus 90s. – tchrist Sep 01 '19 at 13:14
  • @Pablo I heard Portuguese people half-jokingly saying that Portuguese language evolved from Castilian by changing the pronunciation so that Spanish people could not understand. Are you referring to spoken or written Portuguese? – bli Sep 06 '19 at 14:30
  • I live in a Catalan-spoken place, and this claim is absurd. Catalan is in the same family as Occitan which is a language from the South of France. As a native French speaker, without any knowledge of Catalan, I could understand the general meaning of a text, or what people meant when talking slowly. That is not the case for Spanish speakers who, from my experience, don't understand someone speaking Catalan, because there is more vocabulary discrepancy in Catalan/Spanish than in the Catalan/French couple. – Boiethios Nov 28 '23 at 14:10
  • The wikipedia article about Catalan states that despite being spoken mostly on the Iberian Peninsula, Catalan has marked differences with the Iberian Romance group (Spanish and Portuguese) in terms of pronunciation, grammar, and especially vocabulary; it shows instead its closest affinity with languages native to France and northern Italy, particularly Occitan and to a lesser extent Gallo-Romance (Franco-Provençal, French, Gallo-Italian). – Boiethios Nov 28 '23 at 14:11