Greek Lexicography and ESV Translation: Comparing Bauer’s and Louw-Nida’s Lexicons

Greek Lexicography and Translation:
Comparing Bauer’s and Louw-Nida’s Lexicons1

by Vern S. Poythress

[Published in the Journal of the Evangelical Theological Society 44/2 (2001) 285-96.  Used with permission.]


Using Bauer’s lexicon and Louw and Nida’s lexicon presents more challenges to Bible translators than one might suppose, especially in areas where Bauer and Louw-Nida travel in different directions.2 The recent appearance of the third English edition of Bauer’s lexicon, based on the sixth German edition,3 offers a fitting opportunity to reassess the relation of the two lexicons to the tasks of NT interpretation and translation.

Louw and Nida designed their lexicon specifically with the goal of aiding Bible translators.4 The organization by semantic domains, as well as the descriptive definitions of meanings, utilizes up-to-date conceptions of semantics.5 The third edition of Bauer’s lexicon shows improvements in semantic description, but the overall organization of the lexicon is still the familiar standard one. One might therefore naively assume that Louw-Nida would offer superior resources in every respect for any kind of Bible translation. But such is not the case. In my own experience working on the English Standard Version, a conservative revision of the Revised Standard Version, I encountered considerable complexities in using the lexicons. These complexities have convinced me that, for some types of translation, Bauer rather than Louw-Nida serves as the best first resource. And exegetes and translators using either lexicon must understand how its strengths and weaknesses affect its use.


Distinctive Design of Louw-Nida

As early reviewers noted, Louw-Nida introduced two outstanding new features: the organization in terms of semantic domains and the description of meanings by specifying semantic features.6

First, consider semantic domains. Rather than listing word entries in alphabetical order, like more conventional dictionaries, Louw-Nida groups together words with similar meanings, that is, word-meanings belonging to a single “semantic domain.” For example, in Louw-Nida the Domain 26, “Psychological Faculties,” includes νοῦς (“the psychological faculty of understanding, reasoning, thinking, …”),καρδία (“the causative source of a person’s psychological life …”), ψυχή (“the essence of life in terms of thinking, willing, and feeling”), συνείδησις (“the psychological faculty which can distinguish between right and wrong”), φρήν (“the psychological faculty of thoughtful planning, …”), and πνεῦμα (“the non-material, psychological faculty which is potentially sensitive and responsive to God”).7 Bringing together these related meanings help alert us to the differences, similarities, and overlaps in meaning among related terms. At the same time, it accents the difference between different meanings of the same word. πνεῦμα, for example, with the meaning “spirit,” belongs together with other words related to psychological faculties. When it takes the meaning “wind,” it belongs together with other meanings for “Physical Events and States,” Domain 14.8

Second, Louw-Nida describes meanings in an exacting way. Older dictionaries, including Bauer’s sixth German edition and second English edition, have usually been content to offer glosses, that is, English language expressions that are the nearest equivalents to the meaning of the original. But glosses, though convenient as a first approximation, have significant drawbacks. The gloss in English may be potentially ambiguous in meaning. So the dictionary ends up offering several glosses that help mutually to define each other’s meaning. But the various glosses may not be perfectly synonymous. Or they may be too narrow or too broad to represent the original meaning with satisfactory accuracy.9 Louw-Nida therefore chooses to offer a more precise, extended description of meaning. This description includes, as far as possible, all the main semantic features, including any notable connotative associations.10

For example, consider the word ἐκλύομαι. The second English edition of Bauer offers the glosses “become weary or slackgive out.”11 “Give out” is potentially ambiguous between at least two meanings, one to “give out” or present a gift, the other to “give out” in the sense of collapsing from weariness. The other main gloss, “become weary,” helps to indicate that only the second sense of the English phrase “give out” is in view.

But there is still a difficulty. “Become weary” is not completely synonymous with “give out.” “Give out” indicates that one is so weary that he cannot continue. By contrast, the expression “become weary” might involve various degrees of weariness. Can the word in Greek denote any of various degrees of weariness? Or is only severe weariness in view? Bauer does not clearly indicate the actual range of meaning. Louw-Nida includes a description as follows: “to become so tired and weary as to give out (possibly even to faint from exhaustion).”12 Louw-Nida makes it clear that, in its judgment, the word is usually confined to the most severe weariness. Louw-Nida then follows the semantic description with possible glosses: “ ‘to become extremely weary, to give out, to faint from exhaustion.’ ”13

The third English edition of Bauer improves on all earlier German editions and English editions by regularly including extended definitions as well as glosses.14 It has therefore now become comparable to Louw-Nida in its specificity. For ἐκλύομαι it says, “be exhausted in strength, become weary, give out.” The expression “be exhausted in strength,” in boldface Roman type, is an extended definition, clearly indicating that extreme weariness is in view. The expressions “become weary,” “give out,” in boldface italics, are glosses that one might use in actual translation. The use of boldface type separates both the extended definition and the glosses from the surrounding examples and from translations of individual passages (which are given in normal [nonbold] italics).

Finally, Louw-Nida has still one more attractive feature. In quite a few cases, Louw-Nida provides some additional discussion in cases where meanings or customs might be misunderstood in other languages. For instance, under the entry 7.41 for “foundation,” Louw-Nida notes,

In some languages it is possible to describe a typical foundation in ancient times as ‘large stones underneath the walls.’ In other languages, however, this may seem to be quite a meaningless type of expression, since foundations are only made secure by driving stakes deep into the ground. Therefore, it may be best to describe the function of a foundation by ‘what keeps the walls firm’ or ‘how the walls are made not to move’ or ‘what goes beneath the walls.’15

This last feature already illustrates the fact that the value of Louw-Nida depends on the kind of use that one has in mind. Reflection on cultural differences is quite useful for translation into new languages and cultures. It is less useful for translation into English or other languages with a long tradition of interaction with the cultures of biblical times. It is least useful when one is engaging in a revision of an already existing translation, particularly a more literal translation such as the RSV.


Limitations in arrangement of information

Despite these strengths, the character of Louw-Nida limits its usefulness in exegesis and translation. To begin with, the arrangement by semantic domains, though useful for extended study of Greek semantics, is of mixed value for practical purposes of exegesis and translation. A translator typically starts with a Greek word and wants to know which of several possible meanings of the word fits the context. For example, Phil 2:16 contains the expression λόγον ζωῆς ἐπέχοντες “holding fast the word of life.” Or does it mean “being alert for the word of life”? To deal with the expression, one wants to know all the possible meanings of the key word πέχω.

Bauer, through its alphabetical organization, provides the meanings all in one place.16 Louw-Nida does not. The translator using Louw-Nida must start with the index, which lists glosses but provides no contextual information. In the case of πέχω, the index of Louw-Nida lists four possible meanings, with four distinct glosses: “a be alert for 27.59,” “b hold firmly to 31.47,” “c watch 24.33,” and “d stay on 85.59.” The translator who wants further information must look in four different places. The organization tempts the translator to neglect this second step and simply pick the gloss that seems best to fit his own ideas of what Paul is saying. But in doing so, he has bypassed all the strengths of Louw-Nida, and has retreated to the process of using glosses.17 Moreover, since the index of Louw-Nida typically provides only one gloss for each distinct meaning, the information is potentially less accurate than the multiple glosses and the extended definitions that Bauer now provides.18


Contextual information

Bauer also provides more information about the contexts in which particular word usages occur. For example, Bauer distinguishes three meanings of πέχω: (1) “to maintain a grasp on someone or someth.[ing]hold fastτινά someone”; (2) “to be mindful or especially observant, hold toward, aim at, intr.[ansitive], … τινί someone”; and (3) “to remain at a place for a period of time, stop, stay, intr.[ansitive].”19 According to Bauer, these three meanings are distinguished contextually by the fact that meaning 1 occurs with an accusative object, meaning 2 occurs with dative object or an object clause, and meaning 3 occurs with no object. In fact, these distinctions match the meanings offered in Louw-Nida. Bauer’s meaning 1 corresponds to Louw-Nida’s meaning b. Meaning 2 in Bauer, with the dative object, corresponds to meaning a in Louw-Nida, while Bauer’s meaning 2 followed by a clause object corresponds to meaning c. Bauer’s meaning 3 corresponds to Louw-Nida’s meaning d. But Louw-Nida nowhere alerts the exegete to the crucial information that different grammatical constructions accompany the meaning differences.

Consider another example, the use of νήρ. Louw-Nida’s index gives three meanings: “a man 9.24,” “b human being 9.1,” and “c husband 10.53.” Louw-Nida’s index entry for ἄνθρωπος offers the same three meanings, with the same three section numbers. In the sections that discuss these meanings, no distinction is drawn between ἀνήρ and ἄνθρωπος. A student of the New Testament with this much information might easily believe thatἀνήρ and ἄνθρωπος are completely synonymous, and that either can freely mean “human being,” “man,” or “husband,” depending only on the subject matter.20

But to believe so would be a mistake. ἄνθρωπος and ἀνήρ are not, in fact, completely synonymous. In many contexts ἄνθρωπος is nearly equivalent to “human being,” and includes both men and women. ἀνήρ includes a component “male” in its meaning. Bauer’s entry under ἀνήρ makes the situation reasonably clear. It offers as meaning 1 “an adult human male, man, husband.”21

Louw-Nida is not completely unaware of the problem here. In the index, under the entry ἀνήρ, “man” is listed as meaning a, the first meaning, while for ἄνθρωπος the meaning “man” is listed second, after the meaning “human being.” This variation in the order of the list is not accidental, but follows the general principle enunciated in Louw-Nida’s Introduction, “For the most part, the most common or ‘unmarked’ meaning is listed first.”22 But how many people are going to notice this subtle difference between the two entries for ἀνήρ and ἄνθρωπος? And even if they do notice, it is impossible for them to obtain from Louw-Nida any further detail about what are the differences between the two words. Louw-Nida provides absolutely no information as to how we are to discern when the more specific meaning “man” as opposed to the general meaning “human being” is contextually appropriate.

Louw-Nida, by not conveniently and consistently providing information about which contexts activate particular word meanings, limits our ability to make crucial exegetical decisions about which meaning belongs to which context.23 It was not designed for this purpose. It was designed not primarily for exegetes, but for practical translators who might look at an existing English translation or a technical commentator for information on exegetical decisions. It is less satisfactory if a New Testament scholar looks to it for aid in making fresh exegetical decisions.


Limitations in citations of examples

Louw-Nida also limits itself by usually providing only one example of each distinct meaning. Consider the word ἀσπάζομαι. In the index Louw-Nida provides three meanings with three distinct glosses: “a greet 33.20,” “b be happy about 25.130,” and “c welcome 34.55.” Suppose we want more information about the meaning “c welcome.” In section 34.55 we read, “to welcome something or someone, with focus upon the initial greeting—‘to welcome, to accept gladly.’ ” Only one verse, Hebrews 11:13, is given as an example. If we have further questions and want to compare this verse with other cases, Louw-Nida will not help.24

Bauer, by contrast, makes it a practice to list many verses. In most cases, it lists all the verses in the New Testament that use the word in question.25 It is therefore easy to see if a particular use is unusual in some way. Underἀσπάζομαι Bauer offers two main meanings.26 Meaning 1 is “to engage in hospitable recognition of another (w.[ith] varying degrees of intimacy), greet, welcome τινά someone.” Many verses are cited. Meaning 2, which is explicitly marked as “fig.[urative] ext.[ension] of [meaning] 1 in ref.[erence] to someth.[ing] intangible,” is described, “to express happiness about the arrival of someth.[ing]welcome, greet.” Only Hebrews 11:13 is cited from the New Testament, though there is a parenthetical list of citations from other Greek literature. By providing this amount of detail, Bauer makes it clear that the use of ἀσπάζομαι in Hebrews 11:13 is somewhat specialized, but not unheard of.27


Scope of the sample

Bauer further increases its value by providing information from “Other Early Christian Literature,” as the full title to the lexicon reminds us. Louw-Nida provides only meanings attested in the New Testament. Having the extra breadth is useful for exegetes, because we have to worry about whether a Greek word that usually has one meaning in the New Testament may on a few occasions take on another meaning attested primarily outside the New Testament.28 ἀσπάζομαι provides an example. Under meaning 1, “greet,” Bauer notes near the end some possible special uses. Meaning 1b, under the basic meaning “greet,” says, “of short friendly visits, ‘look in on’ … Of official visits pay one’s respects to.” It indicates that this special meanings applies only to Acts 18:22; 21:7, and 25:13 in the New Testament, but attests to its occurrence elsewhere in Greek literature. Under meaning 1a it also describes a special use, “Of homage to a king hail, acclaim” occurring only in Mark 15:18, but also attested outside the New Testament. Louw-Nida provides less of this type of information. Under 33.20 Louw-Nida does note that “It is possible that in some contexts ἀσπάζομαι could be interpreted as ‘to visit,’ ” citing Acts 18:22. But that is all. The fullness of Bauer’s information is clearly an advantage for anyone intent on close exegesis.29


Figurative uses

How do the two lexicons fare with metaphors and figurative uses? Because Louw-Nida was written in the context of principles of dynamic equivalent translation, it gives attention to rephrasing metaphors. It alerts translators to the possibility that a figurative use intelligible in Greek and in English may not be workable in some other language. Such things are useful for translations into new situations. But it is a less useful focus for purposes of exegesis and for working with a more literal translation like the Revised Standard Version.

Once again, ἀσπάζομαι “greet,” may serve as an illustration. In the great majority of uses, this verb is used in the context of greeting other human beings. Hebrews 11:13, alone among passages in the New Testament, usesἀσπάζομαι in a figurative sense to speak about “greeting promises,” that is, welcoming them. Bauer, by placing this meaning physically alongside the primary meaning 1, and by marking it as figurative, helps us to see what kind of use it is. Louw-Nida, by contrast, provides no indication that this meaning is figurative, or that it may be built on the more literal sense “greet.” It also shows another peculiarity. The same passage, Hebrews 11:13, is listed in two separate places, with two separate meanings of ἀσπάζομαι. In 34.55 ἀσπάζομαι is assigned the meaning “to welcome something or someone, with focus upon the initial greeting—‘to welcome, to accept gladly.’ ” In 25.130 ἀσπάζομαι has the meaning, “to be happy about something, on the basis that it would prove particularly welcome (thus implying a type of future orientation)—‘to be happy about, to anticipate with pleasure.’ ” Within both of these two entries occurs a sentence referring us to the other entry. “For another interpretation of ἀσπάζομαι in He 11.13, see 34.55.”30

What is odd about the cross references is that the difficulty may be less with interpreting Hebrews 11:13 than with translating it into other languages. We probably do not have here two distinct senses of the wordἀσπάζομαι, but aspects of a single sense. To “welcome” includes the connotation of being “happy about.” Conversely, “to be happy about,” when it includes the fact that the item in question “would prove particularly welcome,” includes the fundamental idea of welcoming. It appears that there is only one meaning here, including both meaning aspects.

Why then did Louw-Nida split this one verse into two separate entries in two separate places in the lexicon? One does not know. But a hint appears in the section 34.55 that discusses the meaning “welcome.” Louw-Nida says, “In some languages, however, it may be difficult to speak of ‘welcoming promises,’ but one can often render this relationship as ‘they were happy to know about what had been promised.’ ”

If we follow the hint of this remark, a possible explanation arises along the following lines. Our English word “welcome” has among its primary connections the idea of welcoming other human beings. To speak in English of “welcoming promises,” though containing the idea of being happy about the promises when one initially receives them, still carries the associations of the picture of “welcoming” people. There is still a bit of figurative color in the expression. A similar collocation of “welcome” with an abstract like “promises” may not work in another language. In that case, one must look for another expression like “be happy about.” Depending on the resources of the target language, the idea of “welcome” or the idea of “be happy about” may be the better match.

But in the semantic domain classification used in Louw-Nida, “welcome” belongs in the semantic domain 34 Association (primarily personal association), whereas “be happy about” belongs in the semantic domain 25, “Attitudes and Emotions.” Actually, the meaning of ἀσπάζομαι includes both aspects. Welcoming involves both affirming a personal association and being happy about the person welcomed. The division into two separate entries is necessary because of decisions about the organization of semantic domains; but it is an artifact of that organization, not an indicator of two separate meanings of ἀσπάζομαι.

One must therefore distinguish carefully between the needs of translators and students of semantic domains on the one hand, and on the other hand the needs of those who study the texture of a single lexicographical item likeἀσπάζομαι. The classification of ἀσπάζομαι into two semantic domains may serve some needs for dealing with the varying resources of target languages. But it is confusing as an internal analysis of the meaning texture of a single Greek word.

One will find other cases in which Louw-Nida superficially appears to have found or invented a new, otherwise unknown meaning for a word. More probably, what actually has happened is that Louw-Nida is so focused on the issue of translating into other languages that it has not always distinguished between a metaphorical and literal use in the original. Rather, it has listed separate meanings when separate translations might be necessary in some target languages.

As an example, consider the word ἅρπα&#xi;. Louw-Nida recognizes a use as a noun, “robber,” and a use as an adjective, “a vicious 20.4,” “b violently greedy 25.25.” The meaning “vicious” is otherwise unattested, either in Bauer, Liddell-Scott-Jones, Lampe, or Mouton-Milligan’s lexicons.31 Where does this apparently new meaning come from?

Bauer offers the gloss “rapacious, ravenous of wolves.” ἅρπα&#xi; is not a very common word. Its meaning seems to be dependent on the much more common cognate verb ἁρπάζω, having to do with seizing and snatching. Thus, “rapacious” is a good gloss for the adjective ἅρπα&#xi;. We can understand meanings having to do with greedily seizing possessions, or robbing. But this does not give us a meaning “vicious.” The hypothetical meaning “vicious” Louw-Nida finds in Matthew 7:15, which is a metaphor. False prophets, Jesus says, may come in sheep’s clothing, but inwardly they are “rapacious wolves.” The idea of wolves seizing their prey is clearly still visible. This is not a new meaning, “vicious,” but another instance of the old meaning, “rapacious.”

Perhaps Louw-Nida judged that the point of the metaphor was not to say that the false prophets will always want literally to seize you or your possessions, but their behavior towards you will be vicious. Possibly that is correct as an interpretation of the import of the whole verse. But it moves beyond the metaphor to the point of the metaphor. The starting point for the metaphor is the rapaciousness of wolves, not their supposed viciousness. One wonders whether Louw-Nida have distinguished sufficiently between the meanings about wolves, which involve the idea of seizing, and inferences about false prophets derived from the metaphor as a whole.

One can see a parallel difficulty with the treatment of the verb ἁρπάζω. Louw-Nida offers as possible meanings the following: “a snatch 18.4,” “b attack 39.49,” “c plunder 57.235,” “d gain control over 37.28.”32 The fourth meaning, “gain control over,” seems out of place. Again, we find nothing fully equivalent to this hypothetical meaning in Liddell-Scott-Jones, Lampe, or Moulton-Milligan (though Liddell-Scott-Jones and Lampe do offer the meaning “overpower”).33 Under the entry 37.28 Louw-Nida more fully explains, “to gain control over by force—‘to gain control over, to seize, to snatch away.’ ” “Seize” and “snatch away” are appropriate glosses for the more familiar sense of ἁρπάζω. But they are odd glosses if offered as equivalents to “gain control over.” So one suspects that something peculiar is taking place here. Louw-Nida cites the Greek of John 10:28 and adds the following explanation:

… ‘no one will seize them from my hand,’ meaning ‘no one will be able to take them away from my control’ Jn 10.28. Though in Jn 10.28 ἁρπάζω would appear to be in a literal context in view of the expression ἐκ τῆς χειρός μου ‘out of my hand,’ nevertheless ἁρπάζω is certainly figurative in meaning and so is χείρ ‘hand.’

Once again, a metaphorical usage has driven Louw-Nida to produce a distinct lexicographical entry. In the context in John 10, Jesus compares himself to a shepherd and compares those who follow him to sheep. “Out of my hand” is a Semitic stock phrase for “out of my control.” But, in the literal picture on which the metaphor builds, ἁρπάζω has its normal meaning, “seize, snatch.” Wolves will not succeed is seizing or snatching away the sheep. By analogy, attackers will not succeed in removing people from Jesus’ care. By glossing with “seize, snatch away,” Louw-Nida tacitly concedes that it still needs the ordinary meaning “seize” in order for the metaphor to work. It then adds the meaning “gain control over,” not because that is literally the meaning of the word ἁρπάζω, but because such a rendering is appropriate as an interpretation of the point of the metaphor, once one has decoded it and moved beyond the image of wolves and sheep. But to introduce a new lexical entry on the basis of a metaphor like this may confuse students into thinking that we have a new base meaning, not just a creative metaphorical use of a normal meaning.

One may complain in a similar fashion about the meaning b that Louw-Nida offers for ἁρπάζω. Meaning b is “attack.” This hypothetical meaning “attack” is in the vicinity of the normal meaning “seize.” But one does not find the meaning “attack” in Liddell-Scott-Jones, Lampe, or Moulton-Milligan. When one looks under the corresponding section in Louw-Nida, section 39.49, one finds the following: “to attack, with the implication of seizing—‘to attack, to seize.’ ” This does not look like a meaning distinct from “seize.” The example given is John 10:12, translated “ ‘so the wolf attacks (the sheep) and scatters them.’ ” But ἁρπάζω normally has a more specific meaning. The wolf seizes the sheep and scatters them. Of course, a single wolf would not be able to seize all the sheep at once. But he could seize one and kill it, which would be enough to induce others to scatter. Louw-Nida is probably less moved by the improbability of a single wolf seizing all the sheep at once than by the metaphorical character of the passage. As a wolf seizes a sheep, so a evil person attacks the flock of God’s people. In the reality to which the metaphor points, an attack rather than literal seizing of the body is probably the main point. But again, it is confusing to introduce hypothetical new meanings to a word on the basis of a clearly metaphorical context. By similar reasoning one would have to say, in the parable of the lost sheep in Luke 15:3-7, that “sheep” means “one of God’s people,” “lost” means “gone astray from God,” “house” means “heaven,” and “friends” means “angels.” These are not really distinct meanings of Greek words, but a distinct usage in the context of an extended metaphor.

Thus, in this case Louw-Nida is confusing because it classifies a metaphorical use as a distinct new meaning.

The new third of edition of Bauer, like the earlier editions, has only two distinct meanings for ἁρπάζω, not four. But, unlike all earlier editions, it has added “extended definitions.” Under meaning 1 it says, “to make off w.[ith] someone’s property by attacking or seizing, steal, carry off, drag away.” The word “attack” is new in this edition, with no precedent in earlier German or English additions. Likewise under meaning 2, “to grab or seize so as to remove or gain control, snatch/take away,” the expression “gain control” is new and without precedent. The third edition added these expressions in the process of introducing its “extended definitions.” And from where, then, did these extended definitions come? They did not from earlier editions of Bauer. It is hard to avoid the impression that the exact words were lifted from Louw-Nida.

The extended definitions in the third English edition of Bauer appear to offer something clearer and more exact than mere glosses. But are they in fact more exact? If they are merely imported from Louw-Nida, they bring into the orbit of Bauer the confusions about metaphor with which Louw-Nida is afflicted.

So let us look again at the exact wording in Bauer’s third edition: “to make off w.[ith] someone’s property by attacking or seizing.” “Attacking or seizing” suggests two alternate modes by which the property may be taken away. But all the individual texts that Bauer includes under this entry offer examples of wild animals seizing or carrying off prey, and people seizing and carrying off others’ property. None of these examples involve an attack without seizing. The animals and human beings involved must grab hold or seize items in order effectively to tear them or carry them off. Thus “attack or seize” offers a false dichotomy. Every instance involves seizing. The word “attack” should simply be eliminated. Bauer’s extended definition is less exact than it might be, because it has uncritically taken over the inexactness in Louw-Nida.

One must not be too hard on Louw-Nida. One may accept the fact that Louw-Nida, and Bauer as well, discuss separately the prominent “stock usages” of the New Testament, such as the use of “sheep” as a metaphor for people and “shepherd” as a metaphor for leaders of God’s people. One can appreciate that metaphorical usages of all kinds may need especially careful treatment in the process of translation into new languages. Hence, it is understandable that Louw-Nida includes such information. But though obvious uses like “sheep” for “people” are marked as figurative,34 others, such as apparently new meanings of ἁρπάζω, are not. Louw-Nida sometimes generates new senses as an artifact of the desire to devote special discussion to metaphors. And now the new third edition of Bauer, when it takes over its extended definitions from Louw-Nida, introduces potential confusion about the relation between metaphorical and nonmetaphorical uses. But we should note that, unlike Louw-Nida, Bauer’s lexicon appears generally to have added descriptions to existing senses rather than postulate new senses.

More broadly, Bauer itself inevitably suffers at least some limitations in comparison to Liddell-Scott-Jones’ lexicon. We earlier noted that Louw-Nida limited exegetical possibilities by only considering senses that occur within the Greek New Testament. Bauer expands to include “other early Christian literature.” But it still encompasses only a limited selection of Greek. It also includes valuable bibliography of secondary scholarly discussions of word meanings. But it still sometimes fails to include senses that are relevant for NT interpretation. For example, for the entry on ἐπέχω Bauer includes the sense “hold fast,” but not the sense “hold out,” which is attested in Liddell-Scott-Jones and is relevant for the interpretation of Phil 2:16.35 Bauer’s omission occurs merely because ἐπέχω does not have this sense elsewhere in the limited corpus of early Christian literature. But even Liddell-Scott-Jones may on occasion be too limited, now that we have available virtually the full corpus of ancient Greek literature in the Thesaurus Linguae Graecae. In the long run, computerized searches through the TLG may lead to refinement of our lexicography all down the line.



What do we conclude? Louw-Nida may help the translator who is wrestling with conveying metaphors effectively. It will not help the exegete who needs exact information about distinct meanings, uncluttered with an artificial multiplication of senses generated by metaphorical uses.

Thus Bauer is the main and indispensable lexicon to use for serious exegesis of the New Testament. But the exegete must also have an eye on Liddell-Scott-Jones, so as not to miss possible senses that Bauer does not list. And Liddell-Scott-Jones is itself subject to refinement because of the mass of material now available in the Thesaurus Linguae Graecae. Louw-Nida, though profoundly stimulating in various respects, was not really designed for use in careful exegesis, and is likely to be misused by those who try to use it for this purpose. The translator who has finished his exegesis, and who is dealing with a knotty problem with a new language and culture, may look to Louw-Nida for help in conveying the meaning into the new cultural situation.



1 An earlier version of this article was presented under the title “Greek Lexicography and Translation Principles and Purposes,” as a paper at the annual meeting of the Evangelical Theological Society, Nov. 17-19, 1999, Sheraton Ferncroft Resort, 50 Ferncroft Road, Danvers, MA 01923.

2 Walter Bauer, A Greek-English Lexicon of the New Testament and Other Early Christian Literature, rev. and ed. by Frederick William Danker (3d ed.; Chicago/London: University of Chicago Press, 2000) (henceforth BDAG); Johannes P. Louw and Eugene A. Nida, eds., Greek-English Lexicon of the New Testament Based on Semantic Domains (New York: United Bible Societies, 1988) (henceforth, LN). The fact that Bauer’s lexicon has been translated from German creates additional complexities that we cannot explore here.

3 Walter Bauer, Griechischen-deutsches Wörterbuch zu den Schriften des Neuen Testaments und der frühchristlichen Literatur, ed. Kurt Aland and Barbara Aland (Berlin/New York: de Gruyter, 1988). But, as we shall see, in adding extended definitions the third English edition appears to depend primarily on Louw-Nida rather than the sixth German edition of Bauer.

4 “This Greek New Testament lexicon based on semantic domains has been designed primarily for translators of the New Testament in various languages” (LN iv).

5 See LN viii-xi.

6 Early reviews include David Alan Black, Review, Filologia Neotestamentaria 1 (1988) 217-18;Hendrikus Boers, Review, JBL 108 (1989) 705-7; James K. Elliott, Review, NovT 31/4 (1989) 379-80; Francis I. Gignac, Review, CBQ 53 (1991) 334-336; Kenneth Grayston, Review, JTS 41 N.S. (1990) 198-201; I. Howard Marshall, Review, EQ 62 (1990) 183-186; James M. Reese, Review, BTB 18 (1988) 150-51; Moisés Silva, Review, WTJ 51 (1989) 163-67; Nikolaus Walter, Review, Theologische Literaturzeitung 114/11 (1989) 817-20; John J. Welch, Review, Biblica 70/3 (1989) 438-442. As might be expected, given the originality of LN, the early reviews focused for the most part on describing the innovative features of LN. Given the obviously helpful and stimulating nature of these innovations, negative comments were muted.

7 LN 26.1-16 and ix-x.

8 See LN ix, 14.4.

9 For further difficulties with the practice of offering glosses, see LN viii.

10 LN vii.

11 BAGD 243.

12 LN 23.79.

13 Ibid.

14 See BDAG viii for discussion. Earlier editions of Bauer did include extended definitions on occasion, but not nearly as regularly as the third English edition.

15 LN 7.41.

16 “The convenience of a dictionary in the traditional format can hardly be overestimated” (Gignac, Review, 335).

17 LN itself warns, “It would be a mistake, however, to consider that the glosses employed [in the index] in identifying the various meanings are adequate to determine what a lexical unit may mean in a particular context. Only by carefully reading the entry, and hopefully those entries which immediately precede and follow, can one fully appreciate the referential range of any meaning” (LN xi).

18 The early reviews of LN could not anticipate that LN but not BDAG would become widely available in computer readable form. The availability of LN within Bible software programs, combined with the nonavailability of BDAG, tempts NT scholars to use LN as a shortcut.

19 BDAG 362.

20 LN 10.53, in discussing the meaning “husband,” does usefully note that this distinct meaning is “normally clearly marked by context, usually involving a so-called ‘possessive marker.’ ”

21 BDAG 79. Even Bauer shows some imprecision when it discusses meaning 2, in which ἀνήρ is “equiv.[alent] to τἰς, someone, a person.” Most of the instances under this meaning in fact involve male examples, not simply neutral cases of “someone.” (Even Romans 4:8, which expresses a general principle using ἀνήρ, is influenced by the LXX, which may have chosen ἀνήρ because David, a man, is the prime example of the general principle that people may receive forgiveness of sins.) See Vern S. Poythress and Wayne A. Grudem, The Gender-Neutral Bible Controversy: Muting the Masculinity of God’s Words (Nashville: Broadman and Holman, 2000), appendix 2.

22 LN vii.

23 As Silva (Review, pp. 166-67) says, “Another method [of definition] which could have been put to good use is one that stresses collocation, i.e., the distinctive syntactical patterns of words. The better traditional dictionaries (including BAGD) already make limited use of this technique, and so it is surprising that Louw and Nida did not try to exploit it.”

24 “From a pedagogical point of view, at least two examples are normally necessary to avoid ambiguities” (Silva, Review, p. 166).

25 “Students can count on completeness of citation of all except the most common words appearing in the main text of the 27th edition of Nestle” (BDAG x).

26 BAGD 116-117.

27 As Silva, Review, observes, “… frequency of occurrence is a significant factor in understanding the makeup of the vocabulary; including representative references (or at least providing some statistical guidance) should be a high priority for future editions [of LN]” (p. 166).

28 “One could also wish for the integration of data from the larger context of Classical and Hellenistic Greek” (ibid. 167). Silva (ibid.) also points out that this deficiency also affects the value of the organization into semantic domains. The domains should ideally include other terms available in Hellenistic Greek, even if these are not used in the NT, in order to show the contrasts with those that are used.

29 In reviewing LN, Marshall, Review, observes,

However, their work [LN] should in no way be regarded as a rival to or a substitute for the latter [BAGD]. BAGD gives a fantastic amount of detail regarding the forms of words and their occurrences in Greek literature of all periods, and it is often tantamount to a concordance in its full listing of the NT occurrences of words; it also provides bibliographical information. It remains quite indispensable for detailed NT study. The better assessment of LN is that it is complementary to BAGD … (pp. 184-85).

30 LN 25.130.

31 Henry George Liddell, Robert Scott, and Henry Stuart Jones, eds. A Greek-English Lexicon (Oxford: Oxford University Press, 1996); G. W. H. Lampe, A Patristic Greek Lexicon (Oxford: Oxford University Press, 1961); James Hope Moulton and George Milligan, The Vocabulary of the Greek New Testament (London: Hodder and Stoughton, 1930).

32 LN 2:35.

33 “Overpower” involves the idea of physical engagement, and so is not synonymous with “gain control over.” BDAG offers as meaning 2 “to grab or seize suddenly so as to remove or gain control, snatch/take away” (p. 134). Interestingly, the crucial addition “gain control” does not appear in the second English edition (BAGD), nor is there any equivalent to it in the sixth German edition. See below.

34 LN 11.30.

35 LSJ ἐπέχω meaning II.1. See J. B. Lightfoot, Saint Paul’s Epistle to the Philippians (reprint; Grand Rapids: Zondervan, 1953), 118. But in a forthcoming article I argue, on the basis of a limited search of the TLG, that the material in LSJ on ἐπέχω, under the meaning “hold out,” needs reconsideration.