Basic Word Order

What is basic word order?

Although syntax is concerned with the order of all morphemes, basic word order normally refers more specifically to the order of three constituents: the subject, verb, and object. There are 6 possible ways these constituents can be ordered (with the percentages indicating the frequency of the basic order in 1,187 of the world’s languages according to wals.info): SOV (47.5%), SVO (41.1%), VSO (8%), VOS (2.1%), OVS (.9%), and OSV (.3%). Although all orders are attested, human language is certainly not without its preferences: 88.9% of the languages sampled have the subject before the verb. While some languages have rigid word orders, other languages allow a fair amount of flexibility. English is an example with rigid word order, since changing the order of the words results in ungrammaticality, as illustrated with the following paradigm:

• (134) a. *The disciple the fish is catching. (SOV)
• b. The disciple is catching the fish. (SVO)
• c. Is the disciple catching the fish? (VSO)
• d. *Is the fish catching the disciple? (VOS)
• e. *The fish is catching the disciple. (OVS)
• f. *The fish the disciple is catching. (OSV)

Although the orders in d and e are grammatical, they cannot be interpreted with the fish as the object. Thus, only two orders are allowed in English (SVO and VSO), and VSO is only possible as a question and if only part of the verb moves before the subject. In contrast, examples in any order can be found in the Greek of the New Testament, as given by the examples from Kirk (2012):

• (135) a. Ἀβραὰμ ἐγέννησεν τὸν Ἰσαάκ (SVO) Abraham fathered Isaac (Matt 1:2)
• b. Μεγαλύνει ἡ ψυχή μου τὸν κύριον (VSO)
• My soul exalts the Lord (Luke 1:46)
• c. Αἱ ἀλώπεκες φωλεοὺς ἔχουσιν (SOV)
• The foxes have dens (Matt 8:20)
• d. Τούτους τοὺς δώδεκα ἀπέστειλεν ὁ Ἰησοῦς (OVS)
• Jesus sent out these twelve (Matt 10:5)
• e. Ἔπεμψεν φίλους ὁ ἑκατοντάρχης (VOS)
• The centurion sent friends (Luke 7:6)
• f. Ταῦτα γὰρ πάντα τὰ ἔθνη τοῦ κόσμου ἐπιζητοῦσιν (OSV)
• For all the nations of the world seek these things (Luke 12:30)

Greek has a flexible word order, allowing all 6 possible orders. However, this does not mean that all orders are possible in any context. Just as VSO is triggered by a certain meaning in English (namely interrogative), some of the attested orders in Greek are also triggered by some meaningful element.

More on “basic” word order and how it is determined

As we have seen, languages vary based on the range of grammatical word orders allowed, and a meaningful element may affect word order. With these two principles, a natural question arises – in languages where multiple orders are allowed, what meaningful elements would cause a change from the basic order? In order to account for these other meaningful elements that might affect word order, we must first determine the order we would expect to find in clauses without a word-order-triggering meaningful component. This is what a language’s “basic” word order is – the order(s) attested where there is no meaning component triggering a different order. This would then be the order from which other orders are derived when other meanings are present, just as SVO is basic in English and VSO is derived from it when the interrogative meaning is present.

To determine basic word order, then, we would want to know all the possible meaning components that might affect word order, so we can look only at clauses that exclude those elements. Siewierska (1988:8) gives a description of what this basic clause should look like given the meaningful elements that may possibly affect word order:

… the term ‘basic order’ is typically identified with the order that occurs in stylistically neutral, independent, indicative clauses with full noun phrase (NP) participants, where the subject is definite agentive and human, the object is a definite semantic patient, and the verb represents an action, not a state or an event.

In Siewierska’s (1988) description, there are several different categories of conditions. There are syntactic categories, such as the verb being in an independent clause in indicative mood and the nouns being definite full noun phrases. In other words, an imperative verb would not count as an example of basic word order in a language (since it is not in indicative mood), nor would sentences where the subject or object is a pronoun, proper noun, or has an indefinite article (none of which are definite full noun phrases), since all of these are known to possibly affect word order in some languages. There are also semantic considerations, such as the subject being agentive and human, the object being a “patient” (which means they change state by virtue of the action; we called this a theme in Exegesis A), and the verb being an “action.” Finally, there is the vague but important category of pragmatic considerations – the clause should be “stylistically neutral.” Some of these categories are known to affect word order more drastically than others. Because we cannot elicit responses for grammaticality in dead languages, we are left with searching for clauses that fit this description as closely as possible. We can look at how this works for the various clauses in (135) which show different orders.

The example in (135) shows SOV order in the clause Αἱ ἀλώπεκες φωλεοὺς ἔχουσιν ‘The foxes have dens.’ One of the more significant criteria which affects word order is that the object should be a “ semantic patient ,” or theme. Although this notion of patient-hood is very difficult to precisely describe, it is well-known that verbs that have objects which are not physically affected often have a different word order, e.g. the plate in Adriana cracked the plate is physically affected (and changed) by the event of cracking. With the verb ἔχω ‘have,’ the object is not a semantic patient, since the state being referred to by the verb does not affect the object. Dens are not physically affected in any way by foxes having them, nor is a house affected if I have one. In fact, studies on NT Greek word order have 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.”

To give another example, the clause in (135) has OVS order. Kirk (2012:42) notes that examples like this often have a demonstrative pronoun as the object, and the object refers to something that has “just … been specified.” This suggests that the clause is not “stylistically neutral,” because the context is causing the information in the sentence to be packaged in a different way. On a more general level, if we can detect a pattern in the kinds of contexts that this order is found in, we can reasonably hypothesize that whatever kind of context we have found is triggering the different word order. It is, therefore, a derived word order and is not “basic.” In order to determine the basic order, then, we must find a clause where no other element is triggering a special order.

What is the basic word order in NT Greek?

Kirk’s (2012) work is devoted to exactly this question. Her method is simple. First, she takes out all the obvious clauses where syntactic or semantic elements may be affecting the order, such as subordinate clauses or those with pronouns as the subject or object. Of the remaining sentences, she looks for sentences that are in pragmatically neutral contexts where the information in the clause would not be presented in a way that would change the word order (see Syntax C for more on how movement can affect meaning and how it does so in Greek). These are known to be contexts where we typically find basic word order in living languages, the most prototypical context being what is called a broad focus question. This is a question like What happened? where no one constituent in the sentence is being questioned and no participant in the event is known. The only possible example Kirk could find in the NT occurs in Luke 1:34–35 where Mary asks the question πῶς ἔσται τοῦτο; ‘how will this be?’ The angel replies πνεῦμα ἅγιον ἐπελεύσεται ἐπὶ σέ, καὶ δύναμις ὑψίστου ἐπισκιάσει σοι ‘The holy spirit will come upon you, and the power of the most high will overshadow you.’ Both of these clauses have SV order, but neither have an object, so they are not the best examples. However, the fact that SV order is found as an answer to a broad focus question suggests that it is likely a neutral order.

Another such context is the introduction to parables. According to Kirk (2012:40–41), all parables in Matthew and Luke are introduced with SVO order. She gives an example from Luke 14:16 where it says Ἄνθρωπός τις ἐποίει δεῖπνον μέγα ‘A certain man made a large dinner.’ Although it would be ideal if the subject was definite, this clause fits most of the criteria we saw in Siewierska’s description of what a clause should look like that would reveal basic word order. This is further support for SVO being a neutral order.

Finally, Kirk (2012) gives situational sentences as another neutral environment. These are sentences in a “backgrounded clause that occurs in a narrative, describing participants’ psychological state” (2012:38). These sentences do not fit all the criteria laid out by Siewierska, particularly because a psychological state is being described. The so-called psych-verbs (or paraphrases of them) are known to affect word order in some languages (see especially Belletti and Rizzi 1988; Landau 2009). Nevertheless, it could still be the case that we have basic word order in the context described by Kirk; the evidence is just weaker. In these contexts, Kirk finds both VSO (Luke 7:16) and SVO (Luke 5:26) clauses, suggesting that both orders are found in neutral contexts. If VSO truly is neutral in examples like Luke 7:16, it would be a second neutral order, which is attested in some languages (see https://wals.info/chapter/81 where 67 languages are given with two primary orders, 13 of which alternate between SVO and VSO).

Raw statistics provide additional support for the view that SVO is neutral. The following table represents Nikolaos Lavidas’s (2015:329) count of clauses in the NT with subjects in nominative case, objects in accusative case, and verbs in indicative mood:

As can be seen, SVO is by far the most frequent order, outnumbering the next two most frequent orders (OVS and VSO) 3 to 1. Statistics like this would not in themselves determine basic order because meaningful elements that we know affect word order, such as questions, are being tallied along with everything else. However, it helps to confirm our syntactic analysis because if SVO were not a neutral order, we would have to come up with some other story for its frequent occurrence. While VSO might be neutral in NT Greek, all the evidence points towards SVO being the dominant neutral order.

Why does basic word order matter?

The conclusion drawn by Kirk is different from many NT scholars and grammarians, and it affects how you read the text. Runge (2010:207), for example, states the following:

Koiné Greek is a verb-prominent language, where the least-marked and most basic order of clause components is for the verb to be placed in the initial position. When other elements are placed in the initial position, such placement is motivated by some pragmatic reason. This claim is based not on statistics, but on the varying effects that are achieved by “fronting” of clause elements.

Notice the logic here. VSO order is claimed to be basic, and the claim is made based on the “varying effects that are achieved by ‘fronting’ of clause elements.” In other words, VSO is basic because moving an element before the verb results in different “effects.” This is most often described as “emphasis” (Runge 2010:192–193). The claim is that something like the subject is “emphasized” when it comes before the verb, but without a strict definition of “emphasis,” there is no way to show that a given SVO clause does or does not have an “emphasized” subject. In other words, the claim is not falsifiable. Moreover, the reasoning can quickly become circular: VSO order is claimed to be the basic order, and the subject in an SVO sentence is seen as “emphasized” because VSO is basic.

Kirk’s (2012) analysis, on the other hand, utilizes contexts without any special meaning or “emphasis” that should elicit basic word order based on known data from modern languages – and SVO order is found in these contexts. VSO may also be found, but the evidence is much weaker. The conclusion from this is that SVO order does not always have an “emphasized” subject, contra Runge and many other NT Greek grammarians. This does not necessarily mean, however, that SVO order never has a special meaning. We have seen that OSV order is also possible, and the O in such cases is most likely marked for some special meaning. This tells us that there is a position in the syntax that is before the subject that may receive a special meaning (see Syntax C). The subject could move to this position even if it starts out as SVO order, which means that there would be two options for SVO – one with a subject with special meaning, and the other with a subject without special meaning. We cannot conclude that SVO is not basic because there are SVO sentences with a subject with a special meaning, but we can conclude that SVO clauses could have a subject with a special meaning. Thus, Runge is correct to point out that elements placed before the verb may (even often) have a special meaning, but his claim is too strong in that such elements (particularly the subject) need not have a special meaning.

As we have said, these theoretical differences between grammarians affect how you read the text. To illustrate with a trivial example, Olmstead (2019:174) states that φωλεοὺς ‘dens’ in Matt 8:20 (from (135) above) is an object that is “fronted for emphasis.” We have no way to tell what the difference in meaning would be between VO and OV order and how the OV order changes the interpretation with the added “emphasis” without a semantic explanation of “emphasis.” I assume that Olmstead’s logic is that VO is the basic order of the verb and object, so a deviation from this order, namely OV, would be a result of the object being “emphasized.” A syntactic-semantic explanation is more plausible in this case, namely that ἔχω is a light verb and allows OV order, as we (following Kirk 2012) suggested above. Thus, there is no “emphasis” in OV order here. Although I disagree with Olmstead’s particular analysis here, he does recognize that OV is not the expected order when many commentators simply ignore that fact (see Syntax C for more discussion on movement and meaning).

Conclusion and final example

I give one more illustration here to show some of the issues involved with a non-trivial text. In D.A. Carson’s commentary on John, he comments on the clause θεὸς ἦν ὁ λόγος ‘the word was God,’ saying “Indeed, the effect of ordering the words this way is to emphasize ‘God’, as if John were saying, ‘and the word was God !’” (1991:117). Notice, this way of understanding the Greek has a meaningful impact. If we were to say that ‘and the word was God’ is a better translation, there would be a meaningful difference in what John was conveying. In other words, word order can be exegetically (and theologically) significant (again, see Syntax C).

I do not want to claim that Carson is correct or incorrect; rather, I want to point out what is needed to substantiate Carson’s claim. In this clause, the verb is a copula, so there is no direct object, making the discussion of basic word order related but not in a straightforward way. We have said that SVO or VSO may be unmarked, but this only applies to transitive clauses with certain types of verbs. Intransitive clauses, copula clauses, and subordinate clauses may all have different basic word orders, i.e. word orders that are not marked with a pragmatic meaning. What we would want to know for copula clauses is when both the subject and predicate are expressed, what orders do not seem to have a special meaning? We would particularly want to see what order we find when the predicate is a bare singular (anarthrous) noun. These clauses would need to be analyzed in each of the contexts to see if the predicate is obviously marked with a “special meaning” of some kind with some orders. If we had an example where predicate.noun-copula-subj (the order in John 1:1) clearly has a marked predicate, we would then want to see if this meaning necessarily occurred every time the order occurred. This would tell us whether the word order always resulted in a different meaning or not. Finally, we would want to see whether the same kind of meaning could be present with different orders, e.g. does the order subj-copula-pred also have a predicate with the same kind of special meaning at times? This might suggest that the “emphasis” could be a product of something in the context outside of word order. Carson does not go into these details, and we cannot expect him to in a commentary. However, the more troubling thing is that these details have not been worked out in the Greek grammars which Carson is drawing upon to determine what is normal or basic in copula clauses. Without this foundation, our exegesis is like a house built upon sand.