Interpretive Arguments, Syntax, & Statistics

I hate arguments that take some form similar:

Interpretation A cannot be correct because Greek word X never (or rarely) appears beside (or in proximity) to Greek word Y.

Regardless of the truth of the statistical information, I’ve never actually seen a situation where the statistical information provides anything that couldn’t be determined or recognized by the realities of semantics or syntax.

PS – I hate the word “syntactical.”

11 thoughts on “Interpretive Arguments, Syntax, & Statistics

Add yours

  1. Hi Mike,

    Your post is so short that it is not possible to form an opinion about it. At least for me it is.
    But the topic is interesting, since corpus linguistics uses statistical evidence and Givon’s pragmatic voice also counts occurrences.
    Thus is would be nice, if you could give some examples.


    1. A good example showed up on another blog a while back…I’ll see if I can dig it up. The author argued that when Jesus said to the man on the cross “I tell you the truth today you will be with me in paradise,” he could not have meant “I tell you the truth today, you will be with me in paradise” but he must have meant “I tell you the truth, today you will be with me in paradise.” His stastical evidence that “today” never appears verbs of speaking as a modifier anywhere else in the NT.

      He’s right, of course. But that interpretation of the sentence has nothing to do with stastics and everything to do the pragmatic reality that Jesus’ speaking on that particular day is both already known and irrelevant information for the other man on the cross. I think Paul Grice came up with a maxim about that…

      Givon is going to only to statistical evidence as a corollary to evidenced that comes for language function. And to that extent, I would accept statistics. But they are only as strong as the functional explanation that precedes them.

      My first example is one of where the interpretation is correct. A great example of where the interpretation is wrong would be this book:

  2. Statistics-based comments on Biblical Greek seem all too often based upon a minimal and therefore meaningless database of Koine Greek texts.

  3. Example? Well, I find assertions in posts to B-Greek quite commonly that such-and-such sense must be involved in a particular instance, based upon a careful survey of the GNT and LXX alone of Hellenistic Greek texts.

  4. I seem to find that most vocabulary/syntax arguments re: the NT text revolve around questions of authorship and authenticity (i.e., textual criticism). In this area just about every argument has turned out either bogus or inadequate, because:

    (a) no statistical analysis has been done on the specific word/question/issue

    (b) no real scientific analysis of personal style has been demonstrated.

    (c) no real knowledge of author vocabulary can be established.

    (d) worst of all, almost every specific case given is logically and fatally flawed.

    (e) no awareness of the methodology required is shown.

    Example(s): arguments about the “non-Johannine” style of the PA (Jn 7:53-8:11), when the words discussed are not even a matter of stylistic choice. Words identified as “Lukan” turn out to actually be “Synoptic” (i.e., Markan or translation-Greek). Words and phrases plainly “Johannine” are ignored.
    Non-Johannine passages elsewhere in John are ignored. The passage is too short to analyze for vocabulary, syntax or style. Real and previously identified author stylisms (i.e., Mark, Luke, John) are ignored. Johannine parallels of phrase and thought are ignored. Key questions, like “Does John’s Gospel know anything of the passage?” are not even asked. Rather than do analysis, most writers just quote Metzger endlessly. The textual data is misstated, ignored, misunderstood, misrepresented, and dishonestly presented to enhance a given position. Classification of vocabulary, syntax, and stylistic features is not even attempted. Magical phrases like “All scholars agree…” are liberally used instead of evidence or argument.

    Just a sample.


  5. On one level I think you’re right. We shouldn’t rule out an interpretation just because it requires words being used in ways we don’t know if they could be used, because for all we know the words can be used that way, and we’ve just never seen them used that way. But isn’t it some evidence against a view if it involves words being used in ways that we have no knowledge of outside the passage that’s in dispute?

    Once you probe a little further, I think you have to recognize this anyway. You want to favor what you call “the realities of syntax and semantics”. But how do we know those realities? We look at how words are used. We study the language empirically by seeing how the language works, and that includes seeing how words can be used in different combination in contrast to ways we have no evidence for ever having occurred. That makes me think that one kind of argument you approve of is entirely dependent on the kinds of arguments you find illegitimate.

    1. Jeremy, when I read you comment, I feel as if you’ve completely missed the point.

      That makes me think that one kind of argument you approve of is entirely dependent on the kinds of arguments you find illegitimate.

      Yes, this is true, but there is nothing in my discussion here (post and comments) that contradicts that. The basic reality with regard to statistics is that no statistic can stand on its own. There *must* be an explanatory cause.A the strength of a statistic is only as strong as the functional explanation for the basis of that statistic. Thus statements like the one in my post are *always* fundamentally wrong because they do not explain. Even your suggestion about statistics being used in the determination of how language is used is limited by this fact. A statistic cannot define usage if there is no explanation. A statistic *can* be used in making a point only to the extend that it is collaborated with an actual explanatory hypothesis.

      1. I’m not sure what you mean by statistics here. I was simply referring to how people use language, which isn’t necessarily because of some underlying theory you can put together to explain it all. People use language in ways that don’t always fit a pattern.

        For example, in his Philippians commentary (on 3:2 in particular) Peter O’Brien points to studies about how blepete (sorry, I don’t have a good Greek font set up right now) functions. It can mean “beware of” when (1) it’s followed by an objective clause with me and the aorist subjective or (2) by the preposition apo. But when it’s followed by a direct object in the accusative case, as at Phil 3:2, it doesn’t mean “beware of” but “consider” or “take note of”. At least in every case but Phil 3:2 it means that. His argument is that Phil 3:2 is probably like the others and thus probably means “consider” or “take note of”, because nowhere else is the term used to mean “beware of” in the particular construction that it has in this verse. That seems like a pretty good inductive argument about language use.

        Then a few days after reading his discussion of that, I read your post, and it sounds as if you want there to be some underlying theory of something in order to show that the language use as it stands is how the language is really used. I’m not sure what the underlying theory is supposed to be. It seems like it’s more than just an explanation of what it means if O’Brien’s claim is correct. It seems like it’s an explanation of how the other proposed reading can’t be right, and this explanation has to be on non-language grounds.

        But that seems to require too much. I have no non-language ground for explaining why blepete doesn’t mean “start a fist-fight with”. The only ground I have is that the word is never, ever used to mean that anywhere else, so why should it mean that here? It doesn’t necessarily get ruled out by the context, as in the example you gave in your comment in response to Karl. But it seems like a pretty decent inductive argument, even if it’s not an infallible logical deduction. There’s always the chance that we’re dealing with the only extant occurrence of a completely different use of a term from all the other occurrences we know of, but the more occurrences of the word the less likely this should seem. So statistics do seem to me to be a pretty likely inductive argument, at least when there are lots of cases to compare with.

        Now maybe that example doesn’t fall under what you’re talking about, but I’d like to know the difference if so, because at this point your claim about the worthlessness of this kind of argument seems to me to cover an awfully lot of pretty decent arguments, even if it also rightly discards a number of genuinely worthless arguments. But maybe I’m still missing some crucial element of what you’re talking about.

      2. Well, its helpful to see a concrete example of what you’re saying.

        My own view of that particular issue in question is that commentators are trying to make far too fine grained distinctions. Take BDAG for an example (and this is generally true for most common verbs like this one). Danker gives eight supposedly different meanings for βλεπω. And yet, most of them can be connect to one of the first couple senses — including both of the possibilities that O’Brien lists here: “consider” and “beware of” are both easily derived from “see” and “look” without any necessary need to draw a distinction.

        To say it plainly that entire paragraph in O’Brien is one of the main reasons I quit reading commentaries except when I absolutely must (which is rarely). Too often they’re exercises in futility making massive discussions about the tiniest issues that really aren’t that different–and (as in this case) make them appear far, far more different and significant than they ever could be. They make me feel like I’m in a second year Greek exegesis class filled with students arguing about which of Dan Wallace’s categories of the genitive a particular noun is.

        As for what I had actually intended with this post. I will admit that O’Brien’s example here is far more nuanced than what I had been thinking.

        Here’s the actual impetus for writing it:

        “Of all the occurrences of “truly I say to you,” this would be the only one where the verb “I say” is modified by an adverb (at least, from what I have found) and makes the possibility “Truly I say to you today” something of an oddity.”

        The modification of “I say” by “today” isn’t an oddity because “I say” is never modified by an adverb in the NT. Rather, It is an oddity because the vast majority of time, telling someone that you’re telling someone something today is semantically superfluous information that contributes nothing to a conversation.

        In terms of my post. The statistical argument has no grounding in causation. It is merely a correlation that results from a function of language itself. That is (in Gricean terms) “Be relevant. Do not make your contribution more informative than is required.”

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Powered by

Up ↑

%d bloggers like this: