James 2:18: my works, my faith, or both?

[Stephen Carlson]

I don't have that product of Steve's, so I can't consult it. If I had to guess, I think Steve is going by the principle of natural information flow (PNIF), not Levinsohn's ordering principle (though they are often congruent). Levinsohn's default ordering principles are more descriptive than normative, and I'm not sure as far as his current thinking goes that he still holds to them exactly as formulated. (Further, in this case, it's not clear which constituent is the marked one when this principle is violated.)

What's a good definition/resource for the principle of natural information flow?

Steve Runge uses it a lot; I would start there. It's also mentioned in Levinsohn 2000:29 n.2, and his formulations of his default ordering principles seems largely to be an operationalization of that principle. I think Runge prefers to reason from the concept rather than Levinsohn's formulated rules.

[Stephen Carlson]

Ἀλλʼ ἐρεῖ τις, Σὺ πίστιν ἔχεις, κἀγὼ ἔργα ἔχω•
δεῖξόν μοι τὴν πίστιν σου χωρὶς τῶν ἔργων,
κἀγώ σοι δείξω ἐκ τῶν ἔργων μου τὴν πίστιν.

As mentioned, a Focus on 'my works' would have resulted in
κἀγώ / σοι / ἐκ τῶν ἔργων μου / δείξω / τὴν πίστιν.

That would have made an appropriate use of Focus since the 'works' or lack thereof is the actual contrast. However, the author chose not to put a syntactial Focus on 'by my works' but instead moved 'by my works' out of the way and out of the way from receiving the natural/default saliency. Since his opponent claimed 'faith', James ended on 'faith', and claimed it, too.

Do you allow for verbs, under certain conditions, to be (what you call) a contextualizing constituent? I'm thinking of something like:
κἀγώ / σοι δείξω / ἐκ τῶν ἔργων μου / τὴν πίστιν.

The notion of showing is common to both statements; what distinguishes them are the participants.

[RandallButh]

Do you allow for verbs, under certain conditions, to be (what you call) a contextualizing constituent? I'm thinking of something like:
κἀγώ / σοι δείξω / ἐκ τῶν ἔργων μου / τὴν πίστιν.

The notion of showing is common to both statements; what distinguishes them are the participants.

Or edited a bit:
κἀγώ / σοι / δείξω / ἐκ τῶν ἔργων μου / τὴν πίστιν.

Simple answer is yes and no. Yes, I allow for pragmatic movement of verbs, even though in a VSO language this may become a neutralization in many cases.

I am happy to see your question because it recognizes that four movements/placements are necessary for the theoretical proposal: 1. the subject pronoun, 2. dative pronoun, 3. verb, and 4. the "by my works". "By my works" would also have been fronted before the default template, with only the object remaining in the default template. Well, four movements happen in languages, but it is simpler to read with only the pronoun front-placements. On the other hand, intonation would make the additional reading clear, distinguishing the scalar Contextualizations from a distinct Focus "by my works".

[Stephen Carlson]

[I am happy to see your question because it recognizes that four movements/placements are necessary for the theoretical proposal: 1. the subject pronoun, 2. dative pronoun, 3. verb, and 4. the "by my works". "By my works" would also have been fronted before the default template, with only the object remaining in the default template. Well, four movements happen in languages, but it is simpler to read with only the pronoun front-placements. On the other hand, intonation would make the additional reading clear, distinguishing the scalar Contextualizations from a distinct Focus "by my works".

Thanks for your comments. I really wish we could hear the text as it was phrased back then, but we're stuck with what we've got: a text.

Yeah, four movements are complicated, but some of them are necessary in any proposed understanding of the information structure. At least 1 and 2 are necessary, probably one more to account for the order ἐκ τῶν ἔργων μου τὴν πίστιν. This leaves the deemphasis and the movement of the verb behind κἀγώ σοι, a complication I readily concede but nonetheless feel is offset by the better fit to the context. In particular, the final constituent χωρὶς τῶν ἔργων in the preceding clause has a natural saliency and contrasts well with a focused ἐκ τῶν ἔργων of the instant clause.

[Iver Larsen]

The reason I'm skeptical is that I don't think these two factors are enough to explain the word order, and in many languages unstressed elements like pronouns often follow different patterns of placement. In other words, to claim that unstressed pronouns shift forward for phonological reasons is by no means a strange claim. It happens in languages other than Greek, including mainland Scandinavian languages with the so-called "Scandinavian object shift," where a unstressed object pronoun is moved ahead of the sentential adverb. Compare (in Swedish): Jag kysste inte Marit ("I didn't kiss Marit") with Jag kysste henne inte ("I didn't kiss her"). Researchers have linked this object shift (and its optionality in Swedish) to phonological reasons.

Thank you for quoting an example that I can relate to. I am curious about this so-called Scandinavian object shift. Do you have a reference for this analysis? I am sceptical because we have the same in Danish: "Jeg kyssede ikke Marit" (I kissed not Marit) and "Jeg kyssede hende ikke" (I kissed her not). But that is not an object shift, at least not in Danish, nor is the placement of the unstressed pronoun optional.

It's an object shift the way the (generative) linguists conceive it (the pronoun is base generated in the position where the noun is and shifts in front of the adverb), and all the literature I've read applies this term to this phenomenon in Danish as well. So it does indeed exist in Danish.

As for references, there are many available on the web, but I will point two whose author discusses the phonological aspects of the object shift, the first a dissertation on the topic and the second an article focused on the phonological explanation:
Mayumi Hosono, "Object shift in the Scandinavian languages : syntax, information structure, and intonation" (Ph.D. diss, Utrecht)
Mayumi Hosono, "Scandinavian Object Shift as the Cause of Downstep" http://project.sol.lu.se/uploads/media/ ... umi_01.pdf

Whether it exists or not is matter of interpretation and your theoretical presuppositions. I have now looked at Hosono's thesis, although I do not have time to study all of it in detail. What I found confusing was the term "shift" which I am used to from historical and comparative linguistics. In generative linguistics, people often talk about movements rather than shifts. But Hosono does say "In almost all of these Scandinavian varieties, a weak, unstressed object pronoun moves across a sentential adverb. This movement phenomenon is called Object Shift OS". He also says "OS has long been one of the most controversial issues in generative syntax." In the footnotes he quotes several disagreements among linguists. So, we are here dealing with a controversial interpretation of linguistic data (within a particular linguistic framework that I do not find very useful) just like we often deal with Greek data in this forum where people have different interpretations based on their different presuppositions, whether linguistic or theological. I hesitate to draw any conclusions about Greek based on a controversial and dubious interpretation of aspects of another language.

Some of the Danish data and statements in his thesis are mistaken, and he clearly does not know the language, which I would not expect (being a Japanese linguist). He is apparently more familiar with Swedish. His statements about Danish are based on a very limited set of predetermined, constructed sentences, elicited by others than himself.

When I said before that such movement was not optional in Danish, it needs some refinement. It is the unstressed pronoun whose placement is not optional, but the other order occurs if the pronoun is stressed, e.g.
1. Han kyssede hende ikke "He kissed her not"
2a. Han kyssede ikke HENDE "He kissed not HER (but someone else)"
2b. HENDE kyssede han ikke. "HER he did not kiss"
2c. Det var ikke HENDE, han kyssede "It was not HER that he kissed"

In 1. the object pronoun is unstressed and therefore occurs before the adverb. In 2a. the pronoun has contrastive stress (not her, but someone else) and therefore occurs after the adverb. The variants under 2 are different ways of saying essentially the same thing, but 1 has a different meaning. Hosono did not think about including example 2. in his test sentences.

It is not only pronouns that behave in this way:
3. Jeg kom her ikke sidste år "I didn't come here last year"
4. Jeg kom ikke HER sidste år "I didn't come HERE last year" (I did not set my foot in this particular place the whole of last year).

The locative adverb "here" behaves the same way, unstressed near the verb before the negation (or temporal adverbs). Stressed after the negation. The unstressed words are not enclitics. If we do want to talk about movement, I would say that the pronoun in 2a and locative adverb in 4 are moved to the right, away from default position, because they are pragmatically emphasized and consequently phonologically stressed. That is the opposite of Greek where a stressed/emphatic pronoun would move to the left like in 2b.

Hosono has a similar example with the reflexive pronoun where he says "Monosyllabic reflexives move, but disyllabic reflexives do not move"
5. Han så sig ikke i spejlet "He saw himself not in the mirror"
6a. Han så ikke sigselv i spejlet. "He saw not himself in the mirror" (as quoted by Hosono)
6b Han så ikke sig selv i spejlet (as written in correct Danish) "He saw not himself even in the mirror."
Sigselv is not a disyllabic reflexive, but the reflexive pronoun strengthened by the always stressed emphaziser "selv" (even) which can occur in many other contexts, e.g. selv ham (even him), selv om (even though). My guess is that Hosono was confused by the English cognate "-self".

I agree that a pronominal object for a verb is often placed close to the verb(root) across languages whether that object is an affix or a separate word. So, I agree that pronouns behave differently than noun phrases which are not pronouns. But it is not caused by phonological reasons, at least not in Danish, and I am sceptical about such a claim for Greek.

From what I understand of your theory, it does not really explain the differing placement of pronouns vis-à-vis nouns (or noun phrases) in Greek, and I don't understand why you seem so hostile to phonological explanations.

My apologies if I came across as hostile to phonological explanations. I agree that phonology is involved, but more as a result than a cause. It is the traditional explanation with the so-called enclitics that I am questioning, including the idea that an "enclitic" pronoun can in some way "strengthen" the verb. I think my theory provides a better explanation, but I am not asking you to accept that, only keep an open mind.

[D Ryan Lowe]

Levinsohn on p. 39 gives a better explanation for the placement of σοι in Jas 2:18, though he does not apply it to this text: "A change in the position of pronominal constituents often occurs when a focal constituent precedes the verb." Neverthless Levinsohn seems to think this rule applies only when the pronoun would end the sentence, rather than seeing it as being more generally applicable.

Actually, I think Levinsohn doesn't apply this rule to Jas. 2:18 because he doesn't consider κἀγώ to be a focal constituent; rather, it's a point of departure. I don't think Levinsohn ever applies the rule regarding points of departure.

[Stephen Carlson]

Actually, I think Levinsohn doesn't apply this rule to Jas. 2:18 because he doesn't consider κἀγώ to be a focal constituent; rather, it's a point of departure. I don't think Levinsohn ever applies the rule regarding points of departure.

That's probably right, but an adverbial καί is usually a focus particle. Some theories of information structure permit a constituent to be both a topic and one of the foci, in the guise of a contrastive topic. And I think κἀγώ (σοι δείξω) is a contrastive topic. With contrastive topics there's going to be a contrastive focus later on.

[D Ryan Lowe]

Actually, I think Levinsohn doesn't apply this rule to Jas. 2:18 because he doesn't consider κἀγώ to be a focal constituent; rather, it's a point of departure. I don't think Levinsohn ever applies the rule regarding points of departure.

That's probably right, but an adverbial καί is usually a focus particle. Some theories of information structure permit a constituent to be both a topic and one of the foci, in the guise of a contrastive topic. And I think κἀγώ (σοι δείξω) is a contrastive topic. With contrastive topics there's going to be a contrastive focus later on.

What makes you think that κἀγώ is adverbial here? Every translation I've looked at translates it as a conjunction, "and I," and not any of the usual adverbial uses: "indeed," "even," "also."

[Stephen Carlson]

What makes you think that κἀγώ is adverbial here? Every translation I've looked at translates it as a conjunction, "and I," and not any of the usual adverbial uses: "indeed," "even," "also."

Sorry about that. Don't know what I was thinking. At any rate, Levinsohn's "point of topic" (his then-preferred term for a topicalizing constituent) can also include a contrastive topic.

[Stephen Carlson]

What I found confusing was the term "shift" which I am used to from historical and comparative linguistics. In generative linguistics, people often talk about movements rather than shifts. But Hosono does say "In almost all of these Scandinavian varieties, a weak, unstressed object pronoun moves across a sentential adverb. This movement phenomenon is called Object Shift OS". He also says "OS has long been one of the most controversial issues in generative syntax."

Thanks for your thoughts on the Danish Object Shift. Don't too hung up by the term "object shift." It is better to think of this terminology as more like proper names (or labels) than descriptions. It is better to understand the pattern of behavior that the label refers to than to try to deduce it from the name alone as it were a description. Of course, if the label is too confusing, there may be reason to adopt another one and contribute to the linguistic tower of Babel. All my point is that pronouns often have different patterns of behavior for placement than their corresponding noun phrases. This is true in many languages, even in your own mother tongue. I simply ask that it is reasonable that this difference applies to Greek as well.