It’s Greek to You and Me.

A decade ago I posted about the ways different languages have of expressing what English-speakers term “Greek to me,” and there was a lively discussion; it seems a good time for a follow-up, and Dan Nosowitz has posted It’s All Greek to You and Me, So What Is It to the Greeks? at Atlas Obscura. It starts with the phrase “It’s Greek to me” and asks “So where did the phrase come from, and why is its sentiment so universal?”

As with far too many linguistic questions like this, there is no definitive answer. One theory ties it to medieval monks. In Western Europe at this time, the predominant written language was Latin, but much of the writing that survived from antiquity was in Greek. The theory holds that these monks, in transcribing and copying their texts, were not necessarily able to read Greek, and would write a phrase next to any Greek text they found: “Graecum est; non legitur.” Translated: “It is Greek; it cannot be read.” […]

Though Greece is nominally part of Europe, its deep ties with the Middle East, North Africa, and the Slavic countries have meant that Greek culture sometimes doesn’t seem fully of a part with Western Europe. The alphabet used there today, called the Euclidean alphabet, was ironed out just after the Peloponnesian War, in around 400 B.C. But there were several versions of the Greek alphabet and language before then, and one of those, it’s generally believed, was used by a Greek colony in southern Italy. That one was adopted by people who inhabited early Rome, and steadily evolved on its own into Latin. By Shakespeare’s time, the Greek alphabet looked like a weird fifth cousin to the Latin alphabet. […]

English is not the only language to rely on Greek as a shorthand for gobbledygook. Spanish, Portuguese, Swedish, Norwegian, Dutch, and Afrikaans do as well. You’ll notice those are all European languages except for Afrikaans, and Afrikaans is Germanic in origin.

Harry Foundalis, a cognitive scientist who studies Greek linguistics, says many Greek people know that in English and other languages, Greek serves as an indecipherable tongue, and many Greek people, especially young ones, speak English anyway, so they’ve encountered it before. “How do we feel about it? We find it funny,” says Foundalis. “Those of us who know it make jokes with it. For example, I’ve noticed that every time I talk to an English-speaking audience and I use the phrase ‘That’s all Greek to me,’ and the audience knows I’m Greek, I get a thunderous laughter as a response. So, the phrase works well for me.”

There are, however, an awful lot of other languages that have some version of this phrase that doesn’t use Greek. Some of these are weird in their own right. What’s up with the Baltic countries, which think Spanish is so impenetrable? Why do the Danish use Volapük, a short-lived Esperanto-type constructed language created by a German in 1880? When a Bulgarian says “Все едно ми говориш на патагонски,” which uses “Patagonian” instead of Greek, what the hell are they talking about? Do they mean some extinct indigenous Chonan language, or Spanish, which is the dominant language there, or Patagonian Welsh, which also apparently exists?

Nosowitz goes on to talk about Chinese, which “happens to be the most common replacement for Greek in the idiom around the world.” Thanks, Terry!

Comments

  1. David Eddyshaw says:

    Patagonian Welsh, which also apparently exists

    I don’t think I know this Nosowitz person. Apparently he exists.

  2. I’m sure most readers know that for the French, it’s either Chinese or Hebrew.
    A third, deeply offensive and rapidly disappearing option is “c’est du petit nègre” which I don’t need to translate.

    On a tangential note, if you speak a language very badly, you speak it “like a Spanish cow” – e.g. “ Je parle anglais comme une vache espagnole”

  3. Graecum est; non legitur
    I am totally going o use it. Will write GE;NL and use in a TL;DR manner

  4. “It’s all Chinese letter” in Russian. According to the legend, it originates from letter received in Moscow from emperor of Ming China in 1618 which remained undeciphered for several decades, because there wasn’t anyone able to read it.

  5. Lars (the original one) says:

    At least we don’t use esperanto or ido, we’d never hear the end of it.

  6. Ксёнѕ Фаўст says:

    Polish has udawać Greka ‘to feign ignorance, play the fool’ but Chinese is the stereotypical unintelligible language (to jakaś chińszczyzna!).

  7. David Eddyshaw says:

    On first principles (never a good guide in such matters) I’d have expected “c’est du petit nègre” to mean not so much that the language was incomprehensible but that it was very bad French. Is that not the case, then?

  8. Stu Clayton says:

    You’re presumably thinking of c’est du adulte non-francais.

  9. David Eddyshaw says:
  10. John Cowan2 says:

    Petit-Nègre was specifically the pidgin used in the French West African army from 1857 until decolonization; it never formed a creole.

    Here’s a directed graph of what languages use what other languages as their stock examples of incomprehensibility. Lots of lovely comments at the associated Language Log post; consider them incorporated here by reference.

  11. AJP Crown says:

    John Cowan2 – linguistic dualism presented by John Cowan1’s evil twin.

  12. John Cowan says:

    When IBM came out with the programming language they called PL/I (pronounced “P. L. One”) in 1965, they prudently trademarked the names PL/2 through PL/100, just in case.

    PL/I is one of the few general-purpose programming languages with a U.S. or international standard that does not have an open-source implementation (an attempt was made to write a front-end for GCC, but nothing has been heard of it for over a decade). It’s just too big and complicated for a hack, and too old and disused for a production programming language. I once tried to provide a compiler for Dibol, the first language I used in a full-time job back in 1976 and the subject of an ANSI standard, by translating it into Cobol, but I ran up against some Cobol limitations that made it impossible.

  13. Lars (the original one) says:

    Twenty years ago I played around with a parser / interpreter for Algol 68 written in Perl, just to mess with people’s minds, but ran out of tuits half way.

  14. John Cowan says:

    Wow. I’ve always loved Algol 68, and there are several open-source compilers and interpreters for both Algol 60 (GNU Marst, Nase, Algol60i (in Python), Racket, Cim (Simula 67)) and Algol 68 (Genie, Algol68toC) now. I wrote a compiler in and to Perl for Douglas Hofstadter’s demonstration languages BlooP (for primitive recursion) and FlooP (for general recursion).

    The big limitation, if I remember rightly, was that Cobol has no standard way of computing the external name of a file on the fly (though it can be specified in an environment variable or equivalent), whereas Dibol file names are string expressions.

  15. Lars (the original one) says:

    68 was fun. I’m sure there were several occasions where I understood the Report’s algorithm to match recursive types, not least why it would always terminate. (It was expressed as a huge system of mutually recursive string rewriting rules). Or was that the Revised Report? The fun game was to express types as graphs of Perl references, with loops of course, and get as close as possible to a one-to-one correspondence between the Perl code and the string rewriting rules.

    Somewhere in a box I have a photocopy of the errata sheet that was in the library where I worked then’s copy of the Report / RR — the main text was online at the time, I think, maybe even OCR’ed, but not the errata.

  16. David Eddyshaw says:

    Seems as good a place as any to mention JC’s Perl interpreter for TRAC.

  17. Stu Clayton says:

    5 or 6 ?

    I got the laff of the week from this aperçu:

    # [Schwartz, 1993] also stated that the accusation that Perl is a write-only language could be avoided by coding with “proper care” #

    When the caring gets tough, the tough get caring.

  18. John Cowan says:

    Both. The van Wijngarten grammar of the R⁰RA68 covered the syntax and the static semantics, whereas the (much clearer) grammar of the R¹RA68 covered the entire language, being in effect an interpreter. If the grammar reduced to non-terminals throughout, the program would follow the dynamic semantics; if not, it would not.

    “A procedure yielding MOID NEST1 routine text [that is, the body of a procedure] is a formal MOID NEST1 declarer followed by a routine token followed by a strong MOID NEST1 unit!”

  19. Stu Clayton says:

    If the grammar reduced to non-terminals throughout

    This is a dark conceit. Maybe “reduced” has some special meaning here ?

  20. John Cowan says:

    I should have said expanded.

    The idea is that in addition to ordinary grammar rules there are meta-rules, and a potentially infinite number of actual rules are generated by expanding out all references to the meta-rules in the ordinary rules. Here’s a trivial example (metarule names are in caps, spaces are not significant):

    Metarules:
    EMPTY =
    TYPE = int | boolean | string | array of TYPE
    LIST of ITEM = ITEM | ITEM, LIST
    SEQUENCE of ITEM = EMPTY | ITEM | ITEM SEQUENCE

    Rules:
    declaration: TYPE identifier <- TYPE literal]
    digit = 0 | 1 | 2 | 3 …
    letter = a | b | c | … | A | B | C | …
    character =
    identifier = letter | letter identifier
    integer literal = digit | digit integer literal
    string literal = ” SEQUENCE of character ”
    boolean literal = true | false
    array of TYPE literal = [ LIST of TYPE literal ]

    TYPE appears in more than one place in the definitions of “declaration” and “array of TYPE literal”, and it is guaranteed to expand to the same thing on both sides. The result is that the grammar can parse “integer x <- 32" but not "integer x <- true", because there is no expanded rule of the form "declaration = integer x <- boolean literal".

    Similarly, if you want to handle a more long-range condition, you can do it like this:

    Metarules:
    RULECHAR = letter | space
    RULECHARS = RULECHAR | RULECHAR RULECHARS

    Rules:
    where RULECHARS is RULECHARS = EMPTY

    Then a rule containing "where THIS is THAT" at the end of the definition will expand to nothing at all if THIS is foo and THAT is also foo, but to "where foo is bar" if THAT is bar, and since "where foo is bar" is a non-terminal with no definition, the grammar will not be able to match any further input.

  21. Alexander says:

    Paul:

    if you speak a language very badly, you speak it “like a Spanish cow” – e.g. “ Je parle anglais comme une vache espagnole”

    Etymologically, that makes sense, actually. It was originally “il parle français comme un basque espagnol” – he speaks French like a Spanish Basque (there being Basque on both sides of the border, and presumably the Spanish Basque usually had a weaker grasp on French than their northern cousins did).

  22. Thanks David Eddyshaw and John Cowan for the correction and additional link. I guess you are right – all uses I can actually remember coming across are about mangled French.

    And I’d never come across the history of the word (which goes to show how well French history is taught to the French …)

  23. Lars (the original one) says:

    One interesting thing about Algol 68 is that the concept of context-dependent coercions is formalized and unambiguous. Where the syntax has ‘statements’ they expand to ‘expressions’ in a STRONG VOID context, which means ‘anything goes, we’ll throw away the result’. While to get a void value in a WEAK context, you have to explicitly write null or something like that (20 years, innit). For instance, the left side of an assignment needs to be WEAK so a reference doesn’t get dereferenced, because assignment assigns a value to / through a reference. (There were five or so levels, I remember STRONG, FIRM, MEEK, WEAK).

    Other languages do the same, but don’t have the language for it. Does a variable name denote the address of a cell or the value in it? Depends on where you write it. You can fiddle around with & and \ and * and stuff, but the semantics of it is defined in pages and pages of text (and you can have flame wars about what it really means), or sometimes ‘defined by what the interpreter actually does’ in the case of Perl.

  24. Etymologically, that makes sense, actually. It was originally “il parle français comme un basque espagnol” – he speaks French like a Spanish Basque

    That’s one possibility, but not the most likely. Wiktionary:

    Selon l’hypothèse la plus plausible, ce serait une déformation de l’expression occitane « parlar coma un gavach espanhòl » (parler comme un gabatch espagnol). Ici, le mot gabatch désigne le montagnard (le travailleur venant des Pyrénées, pour les travaux agricoles saisonniers).

    Mais certains pensent que cette locution est une déformation de parler français comme un Basque l’espagnol, mais celle-ci n’est pas attestée dans les ouvrages anciens. Peut-être a-t-elle simplement été formée à partir de vache, comme l’autre locution être sorcier comme une vache espagnole.

    Enfin, une autre hypothèse a été proposée par Pierre Marie Quitard : II fut un temps où les habitants du nord de l’Espagne, voisins des contrées méridionales de la France, en parlaient usuellement le langage, tandis que les habitants du midi de l’Espagne avaient un idiome différent, et les premiers, faisant allusion à cette différence, disaient dérisoirement de quelqu’un qui commettait des fautes grossières contre le français, qu’il le parlait comme un baxo. Or ce mot baxo, qu’on employait pour désigner un Espagnol du bas pays ou du midi de la Péninsule, et qu’on prononçait baco, fut bientôt changé en vaco (vache), et de là vint la locution proverbiale.

    I have bolded the bit where it says there is no early attestation for the “Basque” form.

  25. Trond Engen says:

    What’s the etymology of gabatch? CNRTL doesn’t know it at all.

  26. John Cowan says:

    Apparently (though I can’t find anything authoritative, it is plausible by the usual sound laws) the true descendant of Latin vasco ‘Basque’ in French was vace, from which vache would be an easy change when vace was no longer understood. The modern form basque is plainly a borrowing from some Iberian language.

    Similarly, I continue to think that prête-moi ta plume began life as prête-moi ta lume: what use is a pen if you can’t see? But asking the Moon for light makes good sense, in which case la lune and mon ami Pierrot are in apposition, and with the loss of lume, alliteration would make for an easy shift. And at least lume is in Godefroy’s dictionary as a variant of lum ‘lumiére’.

  27. Meanwhile the word gabacho entered Spanish as (I understand) a Spanish word for any Frenchman, and then as (I remember) a Mexican and Mexican-American word for an Anglo-American.

  28. Trond Engen says:

    Ga- must have looked like a Gothic prefix to better ethymologists than me, so naturally I want to have a go were they thought better of it and gave up. A hypothetical **gabahts from the same verb as andbahts “servant” would seem to fit, but nobody knows which verb that is, if one at all and not a half-nativization of Celtic or Latin, so I don’t even know what it would mean. “Partner in service”? A **gabauhts could be a “trading partner”, but I don’t think the vowel fits. Gabaurhja “fellow citizen” is even farther removed but has the advantage of being actually attested.

  29. PlasticPaddy says:

    This germano-gallic-occitan mutant could then have been most suitably borrowed back into German as Quatsch, for which learned scholars have no better etymology ☺

Trackbacks

  1. […] Hat considers, after the space of a decade, why people might say a language is so foreign as to be […]

Speak Your Mind

*