Movement and meaning (information structure, topic, and focus)

(525) σου moved to an unordinary position

καὶ στραφεὶς πρὸς τὴν γυναῖκα τῷ Σίμωνι ἔφη· βλέπεις ταύτην τὴν γυναῖκα; εἰσῆλθόν σου εἰς τὴν οἰκίαν (Luke 7:44) And he turned toward the woman and said to Simon, ‘Do you see this woman? I entered into your house …’

The expectation for possessive σου is that it would appear adjacent to τὴν οἰκίαν, whether before or after (i.e. as σου τὴν οἰκίαν or τὴν οἰκίαν σου). The fact that σου appears before the preposition demonstrates that it has to have moved from its original position with τὴν οἰκίαν. Given that the rest of the passage is contrasting what Simon has done with the woman, we may hypothesize that the σου serves to contrast Simon with the woman. Despite this clear syntactic violation of a norm, I was unable to find an English commentary that mentioned this, and no major English translation does anything to reflect the meaning as represented in the Greek.

Let us pretend that English word order worked like Greek word order, and if an element was moved from its original position, the same kind of meaning would be expressed. We could then translate the last clause as ‘I entered your into house …’ This translation would more accurately reflect the Greek text, since the contrast signaled by the movement of σου would also be signaled by the movement of ‘your’. Unfortunately, English does not express the notion of contrast in the same way as Greek, so the translation ‘I entered your into house’ would not convey the correct meaning to English speakers. Fortunately, English does have a grammatical means to express contrast – it is normally phonological stress. The translation ‘I entered into YOUR house …’ (the all caps is meant to signal phonological stress, but bold or italics could also be used) would reflect the meaning of the Greek.

In this section, we discuss and define the concepts of focus and topic (after introducing basic notions of information structure) and how they are expressed in Greek. These categories of meaning fall under the subfield of pragmatics in linguistics, which deals with meaning but not in the same way as semantics (see the introduction and further discussion below). Our task will be to define these notions more precisely than they have been in the field of NT grammatical studies and to show how Greek uses movement to express these notions. An implicit argument made throughout is that translations can (and oftentimes should) reflect the meaning of movement in the Greek text by utilizing English’s grammatical mechanisms to convey the same meanings, i.e. by graphically indicating phonological stress (i.e. bold, italics, caps, etc.).

Basic notions of information structure

Information in a sentence can be packaged in different ways. For example, SYLVIA ran the marathon (where all caps indicates stress), there is an inference that it was Sylvia, and not someone else, who ran the race. In contrast, Sylvia RAN the marathon changes the inference – we now think that Sylvia ran the race as opposed to completing it in some other way, such as walking. Again, Sylvia ran THE marathon changes the meaning – we now assume there was some particularly special marathon that Sylvia ran. In case you couldn’t guess it by now, Sylvia ran the MARATHON also changes the inference – we now infer that Sylvia ran the marathon and not some other kind of race, maybe a 10k happening on the same day. In each of these sentences, the same basic assertion is being made. Sylvia engaged in an event of running a marathon; however, changing the placement of the stress packages this information in different ways. Doing so causes the hearer to make additional inferences about the information. In linguistics, how this information is packaged falls under the category of Information Structure.

In Krifka’s (2007) influential essay on basic notions of Information Structure (IS), he discusses (among other things) three foundational IS concepts: focus, givenness, and topic. Following his outline, the below sections are devoted to these three concepts but with particular reference to how Greek expresses them. Before diving into these concepts, Krifka also briefly defines the notions of common ground, presupposition, and assertion, to which we now turn. The philosopher Robert Stalnaker, one of the pioneers of IS notions, defines common ground (CG) as “a body of information that is presumed to be shared by the parties to a discourse” (2014:2). To illustrate, suppose the two kids Maverick and Sawyer are having a conversation about whose dinosaur is the coolest, Maverick’s T-rex or Sawyer’s Triceratops. If Maverick says My T-rex can eat your triceratops, he is relying upon shared information to make his statement, specifically that Sawyer knows which T-rex is Maverick’s and which triceratops is Sawyer’s. The use of my and your helps to establish which dinos are being discussed. Notice, however, that when Maverick says that his T-rex can eat Sawyer’s triceratops, he is not claiming something about which dino belongs to which individual. This information, that the relevant T-rex belongs to Maverick and the relevant triceratops belongs to Sawyer, is shared information by the two parties in the discourse, that is, it is part of the common ground.

An alternative way to say that the information about who possesses the dinos is part of the CG is to say that this information is presupposed. Maverick’s use of the words my and your does not make an assertion about who is in possession of which dino. This information is presumed to already be in the CG (the set of shared information) by Maverick when he uses the possessive pronouns. The asserted information is the information that is not shared but about which a claim is being made, namely that Maverick’s T-rex can eat Sawyer’s triceratops. In this case, Sawyer might have a difference of opinion about the information being asserted. While presupposed information is presumed to be part of the CG, asserted information updates the CG, as long as both interlocutors accept the truth of the statement. If Sawyer responds with No, he can’t! My triceratops will stab your T-rex to death (which would almost certainly be his response), then the two interlocutors have two different beliefs. Which dino would win the battle is not part of the CG; each boy is making a different assertion, even though they can (implicitly) agree about other points in the CG, such as whose dino is whose.

Now that I have updated the CG between me and you (the reader) with definitions of the interrelated concepts of CG, presupposition, and assertion, we can turn to focus, givenness, and topic. Before turning to these concepts, it is worth pointing out that everything we have said so far is language-independent. Stalnaker calls this fact the “autonomy of pragmatics,” since actions like presupposing and asserting information are independent of “the mechanisms that languages use to serve those functions” (2014:1). Two points must be held in tension, therefore, as we progress through the below IS concepts. First, that humans can package information in a sentence in different ways may very well be a universal fact about human language. At the same time, however, languages will certainly differ in how IS concepts are expressed. Because English uses intonation to express focus does not mean that Greek does. Greek might, but it may also use other means, such as word order. Indeed, this is one of the central claims of this section: word order is used in Greek to convey different notions of IS.

Focus

The term “emphasis” is often used to informally describe something we intuitively know to be focused. It is commonplace to say that Sylvia is emphasized in a sentence like SYLVIA ran a marathon, but what does “emphasis” refer to? Rooth (1992) developed the idea that the meaning of a focus expression could be modelled with alternative semantics. The basic idea is that placing focus on an element introduces alternatives to the focused element and then negates those alternatives. Thus, SYLVIA ran the marathon contains, as we said, an inference that it was Sylvia, and not someone else, who ran the race. This idea can be represented formally by replacing Sylvia with a variable: x ran the marathon. This would then be negated and contrasted with the proposition that Sylvia ran the marathon. Thus, the meaning of the sentence with focus on Sylvia would be a combination of ‘Sylvia ran the marathon’ and ‘x ran the marathon’ (where x ≠ Sylvia and refers to negation of the constituent that immediately follows it). Focusing other elements would alter the set of alternatives that are introduced by changing the variable to another constituent. For example, Sylvia RAN the marathon would be represented as ‘Sylvia ran the marathon’ and ‘Sylvia x the marathon’ (where x ≠ run). The two sentences differ in the set of alternatives conjured up in the interpretation of the utterance even though the assertion, that Sylvia ran the marathon, remains constant.

Before exploring how Greek expresses focus, it is worth relating this definition of focus to questions, focus-sensitive particles, and the notion of CG developed previously. A question like Who ran the marathon is appropriately answered with SYLVIA ran the marathon, but Sylvia RAN the marathon is odd. Such a response feels like the respondent is not answering the question posed by the speaker. Given what we have said about focus and what we intuitively know about questions, this makes sense. The meaning of the question Who ran the marathon can be modelled in a very similar way to the way we have modelled the meaning of focus: it would mean ‘x ran the marathon’ where the speaker does not know the identity of x. In asking a question about the identity of a particular constituent, the speaker introduces all the relevant, possible alternatives for who/what can satisfy the description. When answering such a question, focus must be on the constituent whose identity is being questioned, since it is this constituent whose alternatives were introduced by the questioner. Answers of questions with a WH-word, therefore, require the questioned constituent to be in focus position.

Focus-sensitive particles are words that have a meaning that introduce alternatives. Krifka summarizes the meaning of the three much-studied particles only, also, and even: “In the case of exclusive particles like only, it is stated that the focus denotation is the only one among the alternatives that leads to a true assertion; additive particles like also express the presupposition that the assertion holds for other alternatives; and scalar particles like even presuppose that the denotation of the focus constituent is extreme when compared to other alternatives” (2007:25). Thus, the following paraphrases represent the meaning of these particles.

• (526) Only SYLVIA ran the marathon (and no one else).
• (527) SYLVIA also ran the marathon (in addition to someone else).

(528) Even SYLVIA ran the marathon (and it was not likely for her to do so compared to others).

In every case, alternatives to Sylvia are introduced; hence, these particles are called focus-sensitive. The additional contributions of only, also, and even are to signal what kinds of alternatives we should think of and how the alternatives are related to the main proposition.

Finally, how does our definition of focus relate to the notion of CG with which we introduced this section? If the CG is the set of shared information between interlocutors and it can be updated via assertions, then speakers need a way to make clear what kind of information the hearer should add to the CG. Focus helps us do this. For example, Only SYLVIA ran the marathon tells us to update the CG with information about what Sylvia has done, but it also tells us to update it with what others did not do. Likewise, SYLVIA also ran the marathon presupposes that others ran the marathon as well. Such a statement is typically made following an assertion of someone else running a marathon, e.g. Bill ran the marathon; SYLVIA also ran the marathon. Thus, the additive focus particle also may be used to facilitate discourse cohesion (that is, the sense that the individual clauses in the discourse are connected).

As we turn to Greek, we must remember Stalnaker’s principle of the “autonomy of pragmatics.” English data can help us determine how speakers use something like focus to convey information (and that is why we started with English examples), but it will not tell us how something like focus is expressed in Greek. In English, focus is expressed primarily through intonation: placing phonological stress on a word makes us interpret that element as focused. Because we lack data on the intonation that would have been used when reading Ancient Greek texts, we cannot know if phonological stress conveyed focus. However, we know that some languages have a syntactic position for focus, so Greek may be a language that utilizes a syntactic slot to convey that meaning. To determine whether it does, we would need to find examples where there is variation from the “basic word order” (discussed in Syntax A) and the element that is in the atypical slot appears to be focused from a semantics perspective (i.e. it negates relevant alternatives). The example with which we opened this section in Luke 7:44 suggests that Greek can, indeed, use movement to express the notion of focus. The possessive pronoun moved into an atypical location, and Simon was contrasted with the woman. Such an example suggests that Greek allows elements within a noun phrase to be focused, but it does not tell us how Greek expresses focus within a whole clause.

WH-question and answer pairs are the typical way to determine the position of a focused element within a clause, but we do not have a good example of this in the NT. Although this is a simple kind of context, it is not very common in literature. There is a decent example found in Josephus in an indirect question (I use all caps in my translations to indicate focus):

(529) Focused element in question-answer pair

• …τίνι τρόπῳ τοσοῦτον προύχει κατ’ ἀρετήν… φάμενος, εἰ κλήμασιν ἑπτὰ δεθείη ἀμπέλου (Ant 5.307–8)
• … in what way did he surpass others in strength …
• [Samson] said, if he were bound with seven BRANCHES of a grapevine …

The WH-question in this example is questioning a constituent in dative case. We can tell that the dative κλήμασιν ‘branches’ has moved because it has left behind the genitive that modifies it, namely ἀμπέλου, in its original position (before movement, the order would have been εἰ δεθείη κλήμασιν ἑπτὰ ἀμπέλου). Unfortunately, this example does not demonstrate the relative order of subject and focused element, since the subject is implicit in the passive verb δεθείη ‘he were bound,’ but we do see the focused element move before the verb. Other examples require us to put the focus position before the subject and even before a complementizer:

(530) Focus position before complementizer

• οὐχ ἵνα λυπηθῆτε ἀλλὰ τὴν ἀγάπην ἵνα γνῶτε ἣν ἔχω περισσοτέρως εἰς ὑμᾶς (2 Cor 2:4)
• Not so that you would be sad, but so that you would know the overflowing LOVE which I have for you

The constituent τὴν ἀγάπην is in contrastive focus, which is characterized by the frame x … not y or not x … but y, e.g. Bradly ate cake, not cookies or Bradly did not eat cookies, but cake. The focused element is cake in this instance with a single salient alternative being negated, namely cookies. Greek uses x …καί οὐ y … or οὐ x …ἀλλά y for these constructions, the latter of which is found here.

In 2 Cor 2:4, the element in contrastive focus has moved to a position we would not expect it to be (though movement does not always occur with contrastive focus; see examples below). Ordinarily, we would expect the object to come after the verb γνῶτε, so the expected order would be ἵνα γνῶτε τὴν ἀγάπην ἣν ἔχω. We also know that the subject is normally found after ἵνα in its own clause (numerous examples could be cited, but 1 Cor 16:6 is clear: ἵνα ὑμεῖς με προπέμψητε). This means that the expected order would be ἵνα + subj + verb + obj. Because the object precedes ἵνα in 2 Cor 2:4, it must have also moved past the subject slot.

We conclude that there is a focus position that precedes the subject, verb, and complementizer in Greek. However, this does not mean that every focused element must move to this position. Below are examples from Kirk (2012) that demonstrate that movement may, but need not, occur for focus:

• (531) Preverbal object with additive καί focus
• εἰ ἐμὲ ᾔδειτε, καὶ τὸν πατέρα μου ἂν ᾔδειτε (John 8:19)
• If you knew me, you would know MY FATHER also
• (532) Postverbal object with additive καί focus
• στρέψον αὐτῷ καὶ τὴν ἄλλην (Matt 5:39)
• Turn to him THE OTHER also
• (533) Preverbal object with contrastive focus
• ἔλεος θέλω καὶ οὐ θυσίαν (Matt 9:13)
• I desire MERCY, not sacrifice
• (534) Postverbal adjunct with contrastive focus
• ἵνα δικαιωθῶμεν ἐκ πίστεως Χριστοῦ καὶ οὐκ ἐξ ἔργων νόμου (Gal 2:16)
• So that we might be justified FROM THE CHRIST-FAITH and not from Torah-works

Some of these examples are less than ideal. If we were to elicit examples from native speakers of Ancient Greek, we would want the same sentence frame for each example, and the only variables would be the context (for the different kinds of focus) and the order of the words (i.e. we would have minimal pairs). For example, although both (533) and (534) exhibit contrastive focus constructions (where a specific alternative is negated), Matt 9:13 has an object in focus position (ἔλεος), while Gal 2:16 has an adjunct in focus position (ἐκ πίστεως Χριστοῦ). This matters because a syntactic position for

6 Kirk contrasts contrastive focus and contrastive focus under negation, the latter technically being the not x … but y construction. Both are contrastive focus in that two things are contrasted and one is explicitly negated (see Kirk 2012:108–112 and the literature cited there for more details).

focus would mean that an element has moved from its normal position to another position in the sentence. In the case of Matt 9:13, we would assume the unmarked order to be VO where the object follows the verb. Because ἔλεος precedes θέλω, it seems to be in a position that is not its normal position, and indeed, this is the typical order for contrastive focus. Comparing this to Gal 2:16, the logic of this example not displaying movement is as follows: we know from examples like Matt 9:13 that there is a contrastive focus position before the verb, and since we have an element that is clearly in contrastive focus after the verb in Gal 2:16, we must conclude that movement to the preverbal focus position is not obligatory. Still, it would be ideal to have an example where the same type of constituent, namely the object, is found postverbally in focus position. Indeed, Kirk later quotes Friberg (1982:179) saying that objects (rather than adjuncts) must be fronted when negated in a contrastive focus construction: “When the object itself contains the negative marker, it must be fronted.” Kirk concurs, and in my own reading of the data, I do as well.

We have established that Greek does exhibit movement to a focus position in some cases, but it need not. We have also begun to narrow down the location of the focus position when it moves: it comes before the verb, subject, and complementizer. The additive focus examples above show that focus-sensitive particles (in this case καί) need not trigger movement, and this holds for μόνον ‘only’ as well:

• (535) No movement with μόνον
• καὶ ἐὰν ἀσπάσησθε τοὺς ἀδελφοὺς ὑμῶν μόνον (Matt 5:47)
• And if you greet only YOUR BROTHERS
• (536) Movement with μόνον
• τοῦτο μόνον θέλω μαθεῖν ἀφ’ ὑμῶν (Gal 3:2)
• Only THIS I desire to learn from you

Most occurrences of μόνον have various elided elements making it difficult to determine the relative order of constituents, but it is rare for μόνον to trigger movement. We can, therefore, make the following generalizations about focus movement in the NT, i.e. these are the ways that the pragmatic notion of focus is expressed in NT Greek:

• A focus position appears before the complementizer (which is also preverbal and pre-subject)
• Tentative: the answer to a WH-question exhibits movement to focus position of the questioned WH-element
• The focus-sensitive particles καί and μόνον allow but do not require movement to the focus position
• Contrastive focus constructions require fronting when the negative marker goes with the direct object
• Contrastive focus constructions do not require focus movement for adjuncts

Those interested in a more in-depth discussion of focus in the NT should consult Kirk (2012) for discussion and more examples, but it is worth noting at this point how the approach we have outlined is different from what is commonly found in NT studies. NT Greek grammarians often refer to focus as “new information,” as Runge (2010:271–272) exemplifies:

Virtually every sentence contains some kind of information that is newly asserted. This information represents the reason the speaker is saying what he or she is saying; it is the point of the communication. This information might be brand new to the hearer; it may provide some new connection or relationship to what was already known or established. This new information, which represents the point of the utterance, was referred to as the focus of the clause (see §9.2.5). By definition, focal information is the most important part of the clause. Its importance is based solely upon its contribution to the context, not its position in the clause. Typically, focal information is placed as close to the end of the clause as possible, in Greek, English, and many other languages.

The definition of focus offered by Runge is fundamentally different from that of Krifka (2007) and what is standard in the formal syntax and semantics literature. As Krifka explains, a definition like Runge’s (which is also found in some strands of linguistics) requires a graded notion of “importance” and is not falsifiable unless made more precise. If focus is expressed with intonational stress in English, the definition can also be shown to make the wrong predictions. Krifka (2007:29) provides the following example:

• (537) A: Who stole the cookie, John or Mary? B: JOHN stole the cookie.

Since John is mentioned in the question itself, it cannot be the case that John is new even though it is in focus (as indicated by the stress). Our definition of focus as the introduction of alternatives does explain this example: JOHN stole the cookie means that John, and not someone else like Mary, stole the cookie. It does not mean that John (who is a new person in the discourse) stole the cookie. Of course, there is (almost?) always something new in a particular utterance. Otherwise, the speaker would not be attempting to communicate something. This is the information that is being asserted to be added to the CG by the speaker. This information is not necessarily found at the beginning or end of the sentence, but insofar as the subject and topic positions often refer to participants in the discourse already mentioned (and insofar as they tend to come towards the beginning of the sentence, at least in Greek), we can say that the end of the sentence often contains information that is part of the asserted information. As we have seen, however, focal information, as we have defined it, often comes near the beginning of the sentence, not the end. Moreover, most VSO constructions would have the asserted content precede the topic (normally the subject) when the V contained new information (as is often the case). We conclude that the term focus (as represented by intonational stress in English) should be reserved for cases where alternatives are introduced and not for importance or newness of the information.

Topic

Krifka (2007:37) describes givenness as “the indication that the denotation of an expression is present in the immediate CG content.” As he notes, there are varying degrees to which an element is salient in the context, and this degree often determines what kind of linguistic device is employed to refer to the given element. To give a concrete example, both the pronoun he and the definite article the signal to the addressee that the referent of the word is present in the CG, but they do so in different ways: “Definite articles can be used to indicate whether a denotation is given in a CG in general, whereas clitics and pronouns typically indicate that their denotations are given in the immediate CG” (Krifka 2007:38; emphasis added).

We noted above that some scholars use the term “focus” for new information in the sentence. Following Krifka and for the reasons given above, I call this the comment rather than “focus”. Topic can be defined with the notions of comment and givenness: “The topic constituent identifies the entity or set of entities under which the information expressed in the comment constituent should be stored in the CG content” (Krifka 2007:41). Informally, we can think of the topic as what the sentence is about.

11 This is a very typical informal definition, but it begs the question: what do we mean when we say that the sentence is “about” something? Krifka’s definition allows us to treat this more precisely as the constituent in the CG which is updated with the information in the comment. Admittedly, the term “topic” is not always used this narrowly or precisely. For an overview of how the term is used in the field, see Birner (2013:210–215).

In our example sentence Sylvia ran the marathon, the sentence is about Sylvia, and the new information about her is that she ran the marathon. Since we are storing new information about the topic constituent in the CG, that entity is typically already in the CG itself; hence topics are often high on the givenness scale. Since the grammatical subject normally does not introduce something new, the topic is often the subject. Thus, the following equations are often made in the literature, though they can only be regarded as tendencies and not strict rules:

(538) (Incorrect) generalizations in the literature about topic and comment

• grammatical subject ≈ topic ≈ old information
• verb phrase ≈ comment ≈ new information

The NT data confirms that these are only generalizations, as Kirk’s (2012) example shows:

(539) Topic as something other than grammatical subject

γυνὴ δέ τις ὀνόματι Μάρθα ὑπεδέξατο αὐτόν. καὶ τῇδε ἦν ἀδελφὴ καλουμένη Μαριάμ (Luke 10:38b–39a)

• A certain woman, whose name was Martha, received him as a guest. She had a sister called Mariam

The dative demonstrative τῇδε marks the possessor and is not the grammatical subject of the sentence, but it is the topic, since it refers to a given constituent which the sentence is about, namely Martha. The new information that we are to associate with her (the comment) is that she had a sister called Mariam.

As noted by Kirk, linguists have enumerated different kinds of topics, and these often have their own intonational patterns and/or places in the syntax. Our definition of topic also allows for different subcategories. In a larger discourse, an entity about which new information is being added may be contrasted with another entity about which different information is being added, i.e. contrastive topics are possible:

(540) Contrastive topics

ὑμεῖς ἐκ τῶν κάτω ἐστέ, ἐγὼ ἐκ τῶν ἄνω εἰμί (John 8:23)

• You are from below; I am from above

Both ὑμεῖς and ἐγώ fit the definition of topic provided above. They refer to the entities about which assertions are being made. However, when we juxtapose these two statements, we are also invited to contrast the two. As Kirk suggests, the constituents ἐκ τῶν κάτω and ἐκ τῶν ἄνω also seem to be focused and placed before the verb. They are two alternatives which seem to implicitly negate the other alternative (the addressees being from below means they are not from above, and Jesus being from above means that he is not from below).

Drawing on work done on different kinds of topics in Italian, Kirk suggests that two other topic sub-types might be present in NT Greek: shifting topic and familiar topic. Her definition of shifting topic is “A newly introduced, newly changed or newly returned topic”, and familiar topic is a topic that is “given or accessible from the discourse” (2012:113–114). As an example of a familiar topic, Kirk provides the example given above in (539) where the demonstrative pronoun refers to a highly salient, given entity. In the same passage, she also suggests we find a shifting topic:

(541) Shifting topic

καὶ τῇδε ἦν ἀδελφὴ καλουμένη Μαριάμ, ἣ καὶ παρακαθεσθεῖσα πρὸς τοὺς πόδας τοῦ κυρίου ἤκουεν τὸν λόγον αὐτοῦ. ἡ δὲ Μάρθα περιεσπᾶτο περὶ πολλὴν διακονίαν (Luke 10:39–40a)

• And she had a sister called Mariam, who had sat at the Lord’s feet and was listening to his words. But Martha was concerned about the many preparations

After introducing Mariam, a statement is made about what she is doing in the relative clause beginning with ἥ. The topic then shifts back to Martha to describe what she is doing. This is marked by δέ, which often (but not always) indicates a shifting topic.

As we have said, it is sometimes difficult to distinguish between the different kinds of topics, and this is further complicated by the fact that the subject of the sentence is often the topic. Since we have said that both SVO and VSO are possible neutral word orders, we do not know in an example like Luke 10:40a whether we have SV order because the subject is in a topic position or because the subject is in a neutral SV position. One way to tease apart the subject from the topic is, as we did in (539) above, find examples where the topic constituent is not the grammatical subject. In a similar way, we can distinguish between the different kinds of topics by finding examples where it seems like multiple topics are present. If both constituents have moved past the verb in such examples, we can conclude that there are topic slots before the verb. In one example, Kirk says that we find the order Shifting topic > Focus > Familiar topic:

But I, not from a human, receive the testimony

Because the negator οὐ is found with the prepositional phrase παρὰ ἀνθρώπου and not with the verb, we know that this is an example of constituent negation and contrastive focus. Thus, οὐ παρὰ ἀνθρώπου must be focused, and it has moved to the focus position before the verb. Since ἐγώ refers to Jesus who is a salient individual in the CG but is not the topic immediately preceding the verse, it is a shifting topic (which the δέ also supports). Because ἐγώ occurs before οὐ παρὰ ἀνθρώπου, shifting topics seem to occur before focused elements. Finally, τὴν μαρτυρίαν is a familiar topic, since the entire passage is about various testimonies. It also has moved out of the expected VO order to a position before the verb.

Before concluding this section, it may be helpful to further explicate the logic on why an example like John 5:34 cannot have any neutral elements (besides the verb). If the default order of V and O is VO (at least with full NP direct objects and verbs that are not light verbs – see discussion below), then an OV order must be marked. This forces us to conclude that τὴν μαρτυρίαν is not neutral. Since we also know that focused elements often move preverbally (and we would normally expect an adjunct like παρὰ ἀνθρώπου to be postverbal), the constituent οὐ παρὰ ἀνθρώπου must also have moved, in this case to a focused position because of the negator. Finally, ἐγώ appears first in the clause, and since we know that an unmarked subject often occurs after a focused element (see 2 Cor 2:4), it must have moved past the focused οὐ παρὰ ἀνθρώπου to a higher position. Using this kind of logic, we can use the order we find in other clauses with similar elements to help us determine the relative order of marked constituents.

Kirk (2012) concludes that the order of the topic and focus slots in NT Greek is shifting/contrastive topic > focus > familiar topic > verb. When we see elements before the verb that we do not expect to be there, they could have moved as a shifting/contrastive topic, a focused element, or a familiar topic. When multiple elements are preverbal, they will appear in the order given above (though see the discussion immediately below for a more nuanced position). The Greek data suggests that syntactic movement may trigger the (pragmatic) meanings we provided for topic and focus.

Conclusion and implications

In the above discussion, we focused on how movement may reflect pragmatic meaning, but we cannot forget our initial discussion on word order in Syntax A. It is worth rereading our explanation of why we have OV order in some cases:

In fact, studies on NT Greek word order have already noted that ἔχω is very commonly found in OV order, even though OV order is far less common than VO generally (see the discussion in Kirk 2012:46–47 and the references cited there). Given that the verb ἔχω does not have a semantic patient (and because it could be classified as what is called a “light verb”), the clause in (135) should not be considered an example of “basic order.”

If some verbs like ἔχω and ποιέω allow for OV order more regularly because of their meaning, then two different meaning-based explanations for OV order are plausible, one semantic and the other pragmatic. The semantic explanation would be that certain verbs are semantically “lighter” than others and for whatever reason, Greek allows these verbs to come after the object. In this case, no pragmatic explanation is necessary. The pragmatic reason would be that the O moves before the V as an indicator of focus or a particular kind of topic. What is crucial to remember is that both options may be viable. We cannot necessarily conclude that SOV order is pragmatically marked – it could be marked semantically.

In fact, this explanation is relevant to the example in John 5:34 given above. The verb λαμβάνω ‘get’ is similar to ἔχω and ποιέω in that it is often found in constructions where the semantic “weight” of the verb phrase is on the object rather than the verb itself. This is reasonable considering λαμβάνω is, in many ways, simply the change of state version of the verb ἔχω. John 5:34, then, could have the object “fronted” not because it is a familiar topic, but because the verb λαμβάνω is functioning more like a light verb. Both explanations are plausible in this case.

Our caution, then, is that word order is a complex phenomenon. There are three points that must be stressed. First, topic and focus are broad categories that do not have monolithic meanings. In order to precisely describe the meaning of a clause, we must attempt to describe the kind of topic or focus that is being expressed in the particular verse. Second, many different meaning categories may affect movement besides topic and focus, including the nature of the verb, the nature of the subject, the voice of the verb, the nature of the object, and even the phonological “weight” of the constituents – to name a few of the most common. When considering whether a clause is exhibiting a pragmatic meaning based on the order of the constituents, we are not asking ourselves how the order of the constituents differs from the supposed “basic word order” of the language – we are asking how the order differs from the expected order given the nature of all the elements in the clause. The syntactic order of the constituents in a given clause is a fact, but the reason(s) why we see the order we see may be varied and complex. Sometimes the best explanation is to posit less pragmatic meaning, not more. The third and related point is that we must take extreme care when arguing for pragmatic meaning. These meaning differences are often very subtle, and it is not always easy to determine precisely what the meaning is. Thankfully, Greek syntax provides clues as to the most appropriate pragmatic meaning, since there are cases of movement for topic and focus. However, it is easy for pragmatic explanations to become circular (e.g. an element must be fronted for topic because it is in the front) when we do not have recourse to native speaker intuitions. Humility in our judgments is a prerequisite as we approach the data.