The Greek ballot measure

« previous post | next post »

Below is a guest post by Jason Merchant.

With the eyes of the world on the developments in Greece this week, the exact form of the question that will be put to Greek voters this coming Sunday, 5 July, in the referendum that Prime Minister Tsipras announced this past weekend is of no small importance, and almost every commentator in the past two days has been wondering just what the question would be. Just released here is a photo of the ballot measure.

In my translation:



The material I've marked in blue is in English in the original, written in the Latin alphabet; the names of the two parts of the proposal are translated into Greek in the parentheses.

Next to this text are two choices with boxes: "IT IS NOT APPROVED / NO" and below that, "IT IS APPROVED / YES".

This ballot measure is linguistically remarkable not just for the mixing of alphabets and languages, especially given that "Eurogroup" is not translated and will likely not be immediately understandable to many older Greeks who don't know the Latin alphabet or who haven't been following the negotiations in print, but also for its complexity, and for its opaqueness–since the set of proposals haven't been widely discussed (or even communicated) in their details in the Greek media, it's a fair guess that many will be largely ignorant of the contents of the proposal itself. The question opens with a periphrastic passive; the answer boxes use a different verb in the synthetic passive. I'll leave it to people with more experience designing and using questionnaires to discuss the complexity of the question itself.

One source of reading difficulty that does not afflict the Greek as it does the English translation is the determination of the antecedents of the relative pronouns: the Greek gender and number agreement makes it clear that the relative pronoun in the original which I've translated with two instances of "which" has "plan" as its antecedent, and that the relative clause I've translated beginning with "that comprise…" modifies "parts". The original also isn't forced to extrapose the relative clause modifying "plan"–Greek allows verb-subject word order, which is used in this case (literally "Must be accepted the plan of agreement which…?").

Finally, note that the "no" option (that which is being recommended by the government) appears above the "yes" option, in an interesting markedness reversal. The accompanying article doesn't indicate whether the ballots will vary in whether the "yes" appears first or not.

Above is a guest post by Jason Merchant.

Just for grins, here's what Google Translate does with the text:

"We have to accept the draft agreement, which were submitted by the European Commission, the European Central Bank and the International Monetary Fund to the Eurogroup of 25.06.2015 and consists of two parts, which make up their single proposal? The first document entitled "Reforms for the Completion of the Current Program and Beyond" («Reforms for Integration of the current program and beyond") and the second "Preliminary Debt Sustainability Analysis" («preliminary Debt Sustainability Analysis").

And here is how Bing Translator renders it:

"We have accepted the draft agreement, presented by the European Commission, the European Central Bank and the International Monetary Fund in the Eurogroup of 25.06.2015 and consists of two parts, which make up the single proposal? The first document entitled "Reforms for the Completion of the Current Program and Beyond" ("Reforms to complete the current programme and beyond") and the second "Preliminary Debt Sustainability Analysis" ("Preliminary Debt Sustainability Analysis").


  1. Neal Goldfarb said,

    June 29, 2015 @ 2:19 pm

    I've recently worked on two cases involving Greek governmental and legal documents, and judging from the (translations of) such documents that I've seen, the language of the ballot measure is typical of those genres. If anything, it's easier to understand than some of the documents I've had to read.

    Obviously, it's possible that the lack of clarity results partly from the fact that I'm reading translations — especially since legal translation values literal fidelity to the original more than idiomaticity in the target language. However, one of my colleagues is fluent in Greek, and even she has trouble understanding the untranslated documents.

  2. Eric P Smith said,

    June 29, 2015 @ 4:10 pm

    The modal "Must the plan… be accepted" strikes me as odd for a referendum. Surely the question for the Greek voters is, or should be, not whether the plan must be accepted but whether it shall.

    I felt the same with the Scottish Referendum last year. The question was, "Should Scotland be an independent country?" I find it very hard to assign any meaning to the statement "Scotland should be an independent country." It depended on so many factors, not least the result of the referendum. We ought to have been asked, "Shall Scotland be an independent country?"

  3. Anschel Schaffer-Cohen said,

    June 29, 2015 @ 5:28 pm

    @Neal Goldfarb: Are those governmental and legal documents intended to be read primarily by lawyers, or by the general public? In the US I'd expect internal government documents to be illegibly jargonish, but referendum questions would ideally at least be more straightforward, e.g. "Do you support raising property taxes by 0.1% to fund elementary education?"

  4. John Muccigrosso said,

    June 29, 2015 @ 8:36 pm

    I don't know the answer to this, but isn't it likely that most literate Greeks of any adult age can read the very similar Latin alphabet (admittedly in contrast to most reader of Latin alphabets)? Especially when signs in Greece tend to look like this one:

    Also, isn't it likely (not sure this one is answerable) that most Greeks who are likely to vote are paying attention to the overall issue and know what "oxi" means in this case?

  5. Dimitri Nakassis said,

    June 29, 2015 @ 9:27 pm

    @Eric: The Greek πρέπει has a broader semantic range than Jason's translation communicates. It also has the sense of English should, as in the Scottish referendum question that bothered you.

    I'm much more sanguine than Jason is that older Greeks can read the Latin script; virtually every Greek reads it every single day, and older Greeks like my grandmother were actually taught Latin in elementary school. My Greek is far from perfect and the question seems straightforward to me, although it is certainly written formally.

  6. D.O. said,

    June 29, 2015 @ 9:58 pm

    I don't know any Greek (it's all Greek to me, you see), but why verbs in the possible answers are different? They both seems to be "egkpinetai" so to say.

  7. Neil Dolinger said,

    June 29, 2015 @ 10:31 pm

    Not that I understand it any better, but I believe the pronunciation of the Greek is closer to "egkrinetai", which makes me wonder whether there is any etymological relation of the Greek word to the English word "agree". I realize that the English word comes from Latin "ad" + "gratus", not from Greek, but perhaps they share PIE roots.

  8. Eli Nelson said,

    June 29, 2015 @ 11:42 pm

    @D.O. and Neil Dolinger: A gamma is used to represent the /ŋ/ sound before a kappa or another gamma, so it's usually transliterated as an n. You could use the romanization "enkrinetai."

  9. GH said,

    June 30, 2015 @ 1:13 am

    The Google Translate and Bing translations are pretty impressive, right? Apart from some subject-verb agreement and prepositions, and the word-order inversion for questions, they seem more or less spot-on, and I actually find them a lot more readable than the human translation.

  10. D.O. said,

    June 30, 2015 @ 1:21 am

    Well, yes (or should I say, d'oh!). Enkrinetai it is. PIE root (from Wiktionary) is *krey- meaning to move, shake, fly around; to separate, divide. Same origin as crime.

  11. Vilinthril said,

    June 30, 2015 @ 4:22 am

    Just to chime in a bit:

    * No, the order of answers will not be randomized, oxi will be on top on all ballots.
    * The (Greek) root of enkrinetai is also evident in the English word "crisis", ultimately also stemming from the nominalization of the verb krinein "to choose".

  12. RP said,

    June 30, 2015 @ 4:38 am

    In the UK it is widely felt that people are more likely to vote "yes" in referendums (hence, perhaps, the government's decision to ask whether people wish to stay in the EU rather than whether they wish to leave) – people like to be positive. There is not much evidence for it, though (most recent referendums have resulted in "no"), so the effect, if it exists, might not be strong.

    As for Greece, Paul Mason says: "Even the word no has meaning. Oxi is a big act of resistance in Greece going back to the Italian occupation. Ne is not such a powerful word." ( )

    Although the question may seem complex, this has surely been the biggest issue confronting Greece for a long time, broadly understood. The details may not be known, but the basic decision is whether Greece should bow to its creditors' demands virtually regardless of the social and economic costs, or whether it should resist at the risk of being forced out of the common currency.

  13. AntC said,

    June 30, 2015 @ 5:10 am

    New Zealand has a fine tradition of citizen-initiated referendums, in which the initiator attempts to skew the question; whilst the bureaucracy's (re-)drafters attempt to un-skew it. (Look particularly at the October 1997 example.)

    The resulting mashup leaves you damned if you vote Yes and pilloried if you vote No. Luckily they're advisory only, and Parliament can safely ignore them.

  14. Jim Andrakakis said,

    June 30, 2015 @ 6:01 am

    @John Muccigrosso "isn't it likely that most literate Greeks of any adult age can read the very similar Latin alphabet ?"

    (not a linguist here, just a greek)

    For most (say, 65-70%) Greeks below, say, 50 that's definitely the case. For older people, I would expect the % to be dropping fast. My father is around 70, finished mandatory education in his time (9 years of school) and cannot read Latin characters.

    "isn't it likely (not sure this one is answerable) that most Greeks who are likely to vote are paying attention to the overall issue and know what "oxi" means in this case?"

    Well given the obvious attention by the media I would say… yyyyyysort of. There are as many interpretations of "ohi" (no) and "ne" (yes) as there are commentators. The purpose of the ballot should obviously be to present the question clearly and unambiguously regardless of the information given from the media.

  15. Jim Andrakakis said,

    June 30, 2015 @ 6:09 am

    @RP "The details may not be known, but the basic decision is whether Greece should bow to its creditors' demands virtually regardless of the social and economic costs, or whether it should resist at the risk of being forced out of the common currency."

    I agree with you, and this basic premise should be obvious. But, as I replied above, it's not. There are many members of the government (including no less than the PM himself) that claim that an "ohi" (no) does not carry any danger of being forced out of the eurozone.

    This claim is so obviously false, that they resort to weasel words as politicians do. E.g. "a No does not AUTOMATICALLY mean we are out of the euro" (who cares if it's automatic or not?) or "there is no LEGAL way for a country to be forced out of the eurozone" (legal no, but political yes).

  16. Oliver said,

    June 30, 2015 @ 7:27 am

    @Jim Andrakakis, re "there is no LEGAL way for a country to be forced out of the eurozone"
    Someone from Austria here. It may be politically possible to force Greece to leave the EU (just leaving euro not possible) – but it would send a disastrous Signal…
    Meaning: The EU can not be trusted and the biggest bullies can Change the rules as they like.
    I do not think (and Ido not hope) this is the message the rest of the EU wants to send…

  17. Jason Merchant said,

    June 30, 2015 @ 8:41 am

    @Dimitri Nakassis: Yes, πρέπει (prepi) can often be translated with "should" and I certainly considered doing that here, but was swayed by a discussion with a Greek semanticist who works on modality that the stronger "must" should be preferred here. (I note that the German here uses "muss". And given that θα πρέπει (tha prepi) is weaker than πρέπει (prepi), I'd certainly use "should" for the former. But it could've gone either way, certainly.)

    @John Muccigrosso: As Jim says, the older generation and those living outside the major urban centers are much less likely to read the Latin alphabet, despite the fact that it's used in advertising a lot and even on signage—I remember hearing from an older villager that I should turn at the "stor" sign (STOP, with the final P read as a rho).

  18. J.W. Brewer said,

    June 30, 2015 @ 10:29 am

    Here's a randomly-selected example of a recent ballot question in New York state (from 2013) written in a fairly boring and neutral style (although note the formal-register vibe of "shall"): "The proposed amendment to Article 8, section 5 of the Constitution would extend for ten years, until January 1, 2024, the authority of counties, cities, towns, and villages to exclude from their constitutional debt limits indebtedness contracted for the construction or reconstruction of sewage facilities. Shall the proposed amendment be approved?" The voter's options are then limited to "yes" or "no" (or the option to mark neither or possibly spoil their ballot by marking both options, although in NY I don't think such a "scratch" has a different legal effect from an abstention). Also, here in NY questions only get on the ballot with the approval of the legislature, so you don't have the issue found elsewhere of fights between "outsider" proponents and "insiders" who may not support the proposal over who gets to control the wording of the summary.

    Re the "yes" effect (should the question be staying or going) noted above, I think most US ballot questions are ultimately structured like this, i.e. the question is should the status quo be changed, and "yes" is the pro-change answer and "no" is the pro-status-quo answer. That seems like a fairly neutral approach, not prone to manipulation in any particular case. But obviously there are circumstances where that analysis doesn't work well, as in Greece where the status quo is generally viewed as unsustainable and the underlying question is which flawed alternative to pursue instead. I guess you could view framing the UK question where "yes" is the continue-the-status-quo option as non-manipulative if you thought the social/cultural/political purpose of the vote was not to answer the question "do you want the status quo to change" but "do you approve/ratify the resolution of this issue your elected officials have already made." That would not be a typical use of such a referendum in the U.S. but obviously usage in the U.K. is extremely unusual and ad hoc, so perhaps it wouldn't be illegitimate.

  19. Stephen said,

    June 30, 2015 @ 2:34 pm

    Just a datum. Last week I read this
    and the two linked PDFs on referendum question testing.

    A few things that I noticed:
    – A lot of concern about making the question a *lot* shorter than the Greek one.
    – Parliament wants Yes/No options even though that is not the best way to phrase it, Stay in/Leave seems favoured.
    – Agreeing with J.W. Brewer, people seem to favour having the 'maintain the status quo' option first.

  20. Neil Dolinger said,

    June 30, 2015 @ 3:20 pm

    @ D.O.
    "PIE root (from Wiktionary) is *krey- meaning to move, shake, fly around; to separate, divide. Same origin as crime." Does the source you found this in have the PIE root for Latin "gratum"? I used wiktionary for "agree" but it only went back as far as Latin.

  21. Alex Fink said,

    June 30, 2015 @ 5:54 pm

    @Neil Dolinger: yes, it makes grātus out to be from a participle *gʷr̥Htos of a root *gʷerH- 'welcome, greet, praise'.

    (If you ask me there wasn't much hope for the connection; a Latin /g/ against a Greek /k/, in root-initial position, is already pretty damning.)

  22. phspaelti said,

    June 30, 2015 @ 9:31 pm

    @Stephen: Stay in/Leave seems favoured.

    No. That simply cannot be the question. Accepting the Eurogroup's demands is the question. Leaving the Euro is a (perhaps inevitable) consequence.
    Most voters will likely favor a "have your cake and eat it too" option anyhow, i.e., reject the Eurogroup's demands, and stay in the Euro.

    On Randomzing the ballots:
    That would be insane. This is an election, not a psychology experiment. It should be permissible to print a ballot in a public space outside the poles and voters should expect the ballot to look the same as expected. Voting should not be some kind of literacy/intelligence test.

  23. Eli Nelson said,

    June 30, 2015 @ 9:44 pm

    @phspaelti: I believe the "Stay in/Leave" wording Stephen referred to is about the UK referendum, not the Greek one.

    About randomizing the ballots, I don't know how much literacy/intelligence it takes to distinguish "yes" and "no" (or "ΟΧΙ" and "ΝΑΙ" in this case). Having an equal number of ballots with each possible order of these two is all the randomness that would be introduced.

  24. Aristotle Pagaltzis said,

    June 30, 2015 @ 10:20 pm

    For what it's worth, the untranslated ballot verbiage reads very clearly and plainly to me. (I agree with the demographic breakdown of comprehensibility mentioned in previous comments.) It is stiffened only by the passive construction which avoids any reference to who or what is doing the accepting and approving.

    And thus to the subject of "must" vs "should": I think that question depends on how you interpret the agency. Namely: should the Greek people accept, on one hand; must the Greek government accept (on behalf of the people, who have expressed that this is their will), on the other. I believe both of these meanings are implied. This is possible in Greek due to the should/must ambiguity of "πρέπει", which has no equivalent in English, rendering the text untranslatable in its exact meaning. (I find this to be the case surprisingly often, between any two languages. It is always possible to say a very similar thing in both, but very often impossible to say the same.)

    Also for what it's worth, I might be tempted to forgo grammaticality in English to illustrate the gender agreement by adding an explicit "it": the original is thus literally-of-sorts "Must it be accepted the plan of agreement which…?". This is similar (but converse) to how gender agreement often makes an explicit subject entirely superfluous, such that e.g. the English phrase "it's raining" becomes merely "raining [neuter subject form]" in Greek. Because English does not have gender agreement, a translator is then forced to choose between either adding an explicit subject that was not present in the source material (and may not be grammatical in the target language) or else losing that information.

  25. Keith said,

    July 1, 2015 @ 12:28 am

    There is a very good episode either of "Yes, Minister"" or of "Yes, Prime Minister" (I forget which), in which the two main characters (the permanent private secretary (PPS) and his notional boss, the minister) are talking about an opinion poll. The minister wants to know if the people of the country are in favour of bringing back National Service (one or two years of compulsory military service).

    The PPS explains that the minister needs to decide what his policy is to be, and then phrase the poll question that will get the required response in order to validate the policy choice.

    The phrasing of the Greek referendum ballot paper looks to me like a very deliberate attempt to scrupulously avoid influencing the choice of the voter by being extremely neutral and making absolutely no attempt to present the content of the two documents that together make up the EU proposal.

    Also, I do not find it in any way surprising that the wording of a ballot question (this or any other) is phrased using verbs such as "could", "should", "might" (for want of a better phrase, I'd say that these are not verbs of obligation, but of possibility or counsel), while the voter's response is a strictly unambiguous "agree / yes" or "disagree / no".

    The response is a clear, binary choice to agree or disagree with the statement on the ballot, however it is worded.

    If I was a cynic, however, I might just suggest that this will be seen as an ultimatum from a foreign power, and that Sunday 5 July will be presented as a new Επέτειος του «'Οχι».

  26. Stephen said,

    July 1, 2015 @ 5:14 am


    Eli Nelson is correct, the "Stay in/Leave" wording is in relation to a UK referendum on membership of the EU, not this Greek one on membership of the Euro*.

    There was a private member's bill
    ( )
    in, IIRC, 2013 for such a UK referendum. That did not become law but there is now a promise from the government to hold such a referendum before the end of 2017.

    Because of the private member's bill, the Electoral Commission has been working on the wording of the question (and associated responses) for some time.

    That was why I mentioned this.

    * The Greek referendum might be de jure about the response to the Eurogroup's demands but de facto it is about membership of the Euro.

    A few days ago I saw a reference to opinion polling that agree with the idea that most voters favour the "have your cake and eat it" option, it said that c. 80% supported an end to austerity but that c. 80% also wanted to stay in the Euro.

    So, if there is a Greek referendum perhaps it would be more honest to make it explicitly about staying in or leaving the Euro.

  27. RP said,

    July 1, 2015 @ 12:43 pm

    No one knows the consequences for sure, Stephen. The creditors could react to a "no" by offering improved terms, or they could react by forcing Greece out of the euro. They have certainly given the impression they will do the latter, so presumably Greeks are aware of the risk that that is the case.

  28. Joseph Pentheroudakis said,

    July 1, 2015 @ 5:55 pm

    Pronunciation: 'engrInete' (sorry, no IPA keyboard!). Voiceless stops are voiced after nasals, at least in everyday speech.

    Etymology: the active form is εγκρίνω "(Ι) approve." Κρίνω (´krIno') means '(I) judge, (I) estimate'; no infinitives in MG, so the citation form for verbs is the first person singular, present tense. Check 'crisis' for its PIE reconstruction.

    Explaining the issues to the public: there are going to be nightly discussions and debates all week long on at least one radio/television station that I know of. Greeks are profoundly engaged in the political process, so you can be sure they will have a decent understanding of the issues. Whether those issues will be discussed fairly etc is another matter, of course.

  29. ryanwc said,

    July 2, 2015 @ 10:26 am

    I agree that this is a remarkable thing. But as an election official in a jurisdiction near Chicago, I'd suggest that what is remarkable about it is not that it is remarkable in comparison to other referendums, but that referendums are routinely written inscrutably.

    I console myself with the thought that if you wander into the polling place to read a referendum without having decided how to vote, you may be beyond hope anyway. Rational and irrational voters are unlikely to, and should not, make a decision on a referendum based purely on reading the text. They aren't meant as documents that provide all facts and context necessary for decision-making. So ultimately, most voters do and should walk into the voting booth knowing that they're going to vote "It is approved/Yes" or "It is approved/No" based on research, gossip or the opinion of the person who urged them to vote, and not based on a careful reading of the text alone.

    To make fun of the impenetrability of referendums, I once started a contest in my office for the best referendum haiku. The winner, based on a referendum to dissolve a unit of government responsible for water treatment, locally known as a sanitary district, was:

    Shall the neighborhood
    of Norwood Township be made

    I like to imagine a Japanese artist of the 19th century deciding what watercolor drawing might illustrate such a text.

RSS feed for comments on this post