Criswell Theological Review 3.2 (1989) 353-372.
Copyright © 1989 by The
]ORGIZESQE IN EPHESIANS 4:26:
COMMAND OR CONDITION?1
DANIEL B. WALLACE
It would be very appropriate to develop in this paper something of a
"theology of anger," or, more specifically, a "theology of righteous
human anger." Such a study is sorely needed. But it must be built on
the exegesis of several key passages. Our goal, therefore, is far more
modest: we wish to focus on only one text which, nevertheless, con-
tributes heavily to such a theme. Eph 4:26 is arguably the crux inter-
pretum in the NT regarding the validity of man's dikai<a o]rgh< (as the
Greeks put it)--man's righteous indignation.
Why is this so? How can this one verse be regarded as so crucial
to the issue? It is simply because we have great difficulty finding
explicit statements in the NT in praise of human wrath. (One overly
zealous writer went so far as to use the anger of the king in the
parable of the wedding feast [Matt 22:7] as a proof-text for the
validity of righteous human indignation2--in spite of the fact that
1 This is a revision of a paper read at the annual meeting of the Evangelical
Theological Society held at Gordon-Conwell Seminary (December 5,1981).
2 H. C. Hahn, "Anger, o]rgh<," in The New International Dictionary of New
Testament Theology [=DNTT] (3 vols.; ed. C. Brown; Grand Rapids: Zondervan,
1915) 1.110. Further, even if the king in this parable could be interpreted as representing
man (rather than God), the incidental comment by Jesus of the king's wrath (w]rgi<sqh
in Matt 22:1; o]rgisqei<j; in Luke 14:21) is hardly adequate as proof of his sanction of
human anger, for elsewhere he uses questionable moral models in his parables as an
illustration in a different realm of a good moral virtue (cf. the parable of the workers in
the vineyard [Matt 20:1-16]: he is not advocating that every landowner pay the same
wage to all-day and part-day workers; and the parable of the talents [Matt 25:14-30]:
surely he is not here equating wealth with righteousness [cf. also Luke 16:1-9]. Our
point is simply that the parables do not always have a direct, literal application--often,
if not usually, they are illustrative of a truth in an entirely different realm).
354 CRISWELL THEOLOGICAL REVIEW
most would view the king as representative of God.) Consequently,
the imperative o]rgi<zesqe, "be angry," in Eph 4:26, if taken as a
command, becomes the most explicitly positive statement of human
anger in the NT.
I. Possible Syntactical Nuances for ]Orgi<zesqe in Eph 4:26
That o]rgi<zesqe is a command is by no means a settled issue
among the commentators; in fact, some even doubt that it is an
imperative. Altogether I have found in the commentaries seven differ-
ent syntactical options--five of which treat the form as imperative,
two as indicative:
(1) Declarative indicative: "You are angry, yet do not sin."
(2) Interrogative indicative: "Are you angry? Then do not sin."
(3) Command imperative: "Be angry, and do not sin."
(4) Permissive imperative: "Be angry (if you must), but do not
(5) Conditional imperative: If you are angry, do not sin."
(6) Concessive imperative: “Although you may get angry, do not
(7) Prohibitive imperative: "Do not be angry and do not sin."
In order to make this discussion manageable, we need to pare
down the field. We will do this in two ways: first, three options will be
quickly dismissed since their exegetical bases are tenuous at best;
second, three nuances will be grouped as one because in this passage
there is very little difference among them.
A. Implausible Options
The two approaches which treat o]rgi<zesqe as an indicative and
the one which sees it as a prohibition are implausible on their face. I
have seen but one commentator treat the verb here as a declarative
indicative. R. O. Yeager argues that “o]rgi<zesqe in our verse can be
present middle indicative. Taken with concessive kai> such a transla-
tion makes as good sense [as an imperative] and fits the context
well."3 He translates it, "Although you are provoked, do not go on
sinning," rendering this not materially different from a concessive
3 R. O. Yeager, The Renaissance New Testament (18 vols.;
4 Yeager apparently is uncomfortable with the concessive imperative view: "There
is nothing in the imperative mode itself to imply consent or permission" (ibid.), which
has probably prompted him to attempt to make his view rest on more solid syntactical
ground (since declarative indicatives, unlike concessive imperatives, are common).
Wallace: ]ORGIZESQE IN EPHESIANS 4:26 355
There are three primary5 difficulties with this view however:
(1) o]rgi<zesqe is in the thick of an overtly parenetic section, Eph
4:25-32, being surrounded by ten imperatives and two hortatory sub-
junctives; though there are three indicatives6 here, they all speak of
positive realities which God has effected for the believer and as such
constitute the basis for the parenesis.7 The flow of argument, there-
fore, is decidedly against an indicative o]rgi<zesqe. (2) To treat the kai>
which joins o]rgi<zesqe to mh> a[marta<nete as concessive (or adversative)
is doubtful enough between two imperatives (a]lla> or de> would be
expected), but to consider it as introducing the abrupt shift from
indicative to imperative seems especially unnatural.8 (3) Finally, the
entire clause, o]rgi<zesqe kai> mh> a[marta<nete, exactly reproduces the
LXX rendering of Ps 4:4, where it must be taken as an imperative.9
Whether or not the apostle intentionally alluded to this text is not the
point here: even if he used if rhetorically, it is a supreme case of
petitio principii to view the formal correspondence with the Psalm as
having no effect on the syntax in the Ephesians passage.10 This ap-
proach, therefore, must be judged highly improbable--at best.
The second view, that o]rgi<zesqe is an interrogative indicative
(held by Beza, Meyer, and J. Eadie),11 comes under the same judgment
5 A fourth difficulty (though less significant) also presents itself: Yeager's view
tends to see Ephesians as written to a specific, identifiable situation (for anger is stated
as a present problem in the community), rather than as a circular letter. Attempts to
treat Ephesians as addressed to a specific community with a specific set of prob-
lems/needs have not been entirely persuasive. See later discussion.
6 e]sme>n in v 2.5; e]sfragi<sqhte in v 30; and e]xari<sato in v 32.
7 If o]rgi<zesqe as a declarative indicative were treated the same way, then anger
would be seen as a permanent and positive moral virtue (and one which, incidentally,
believers did not possess before salvation).
8 Further, we would most naturally expect the concession to come at the beginning
of v 26--either implicitly (e.g., instead of o]rgi<zesqe we might expect Paul to have
written o]rgizo<menoi) or explicitly (e.g., kai<per).
9 Although there is doubt over the lexical choice of the LXX translator, he has
correctly rendered the syntax of the Hebrew Qal imperative vzgr.
10 H. A. W. Meyer (Ephesians in MeyerK, 2.54) argues cogently against the inter-
rogative indicative view on the basis of the quote of Ps 4:4: “Against this we cannot
urge--the objection usually taken since the time of Wolf--the kai>, which often in rapid
emotion strikes in with some summons. . . ; but we may urge the fact that Paul
reproduces a passage of the LXX (which, it is true, is quite arbitrarily denied by Beza
and Koppe) in which o]rgiz. is imperative, and that such an abrupt and impassioned
question and answer would not be in keeping with the whole calm and sober tone of
the discourse." Similarly, cf. J. P. Lange, Ephesians in Commentary on the Holy
burgh: T. & T. Clark, 1897]) 140.
11 See MeyerK, 2.52 and J. Eadie, A Commentary on the Greek Text of the Epistle
of Paul to the Ephesians (reprint ed.;
names. The view is no longer popular.
356 CRISWELL THEOLOGICAL REVIEW
for the same reasons: it would be an uneasy indicative in the midst of
imperatives, the allusion to Ps 4:4 shows that an imperative is in Paul's
mind, and the use of kai> in the sense of "then" or "therefore" is not
The third implausible view is the prohibitive imperative view-
i.e., that the negative mh> governs both a[marta<nete and o]rgi<zesqe.13
This view takes a "180-degree" turn from treating o]rgi<zesqe as a
positive injunction. In spite of the theological difficulty caused by the
prima facie reading of "be angry" as a command, this view is impos-
B. Permissive, Conditional, and Concessive Grouped Together
On a popular level especially, the permissive, conditional, and
concessive views are all neatly separated. But several writers hold out
for the distinction, at least, between permissive and conditional.14
Thus, J. L. Boyer states that "in Eph 4:26 it is difficult to understand
'Be angry and sin not' as a command or even a permission, expecially
[sic] in light of the context. . . . It is much easier to take it as a
condition. . ."15
This distinction is usually made because the imperative can have
a permissive or conditional nuance. Grammarians, however, make no
12 Not only would we normally expect a@ra or ou#n here, but in Paul's usage
especially we are accustomed to seeing explicitly paratactic structure if that is what he
13 C. Hodge seems to entertain this view (A Commentary on the Epistle to the
. . . the words of the apostle may mean, do not commit the sin of being angry. To this it is
objected, that it makes the negative qualify both verbs, while it belongs really only to the
latter. It is not necessary to assume that the apostle uses these words in the precise sense
of the original text; for the New Testament writers often give the sense of an Old
Testament passage with a modification of the words, or they use the same words with a
modification of the sense.
14 Aquinas embraced the permissive view; more recently, cf. H. Alford, The
Greek Testament, vol:
3: Galatians-Philemon (3 vols.; rev. E. F. Harrison;
Moody, 1958) 125 (though he calls it "assumptive"); R. P. Martin, Ephesians in The
Broadman Bible Commentary (12 vols.;
seems to lump conditional, concessive and permissive ideas together, his translation
reflects the permissive idea: "You may be angry. ..if you can't help it. . . .”). Others
have held the conditional view, considering it as different from the permissive view: cf.
L. Mitton, Ephesians
(NCB; ed. M. Black;
1973) 168: "It is quite wrong to take it as a command or even a permission to be angry";
J. Gnilka, Der Epheserbrief (in HTKNT) 235, asks, "Wird hier der Zorn fur gewisse
FaIle konzediert?" ("Is the anger allowed here for particular cases?") He answers in the
negative because anger in v 31 is prohibited.
15 J. L. Boyer, "Other Conditional Elements in New Testament Greek," GTJ 4
Wallace: ]ORGIZESQE IN EPHESIANS 4:26 357
distinction between a conditional imperative and a concessive im-
perative.16 And semantically, of course, concession is one kind of
condition. In this context, since o]rgi<zesqe is followed by a prohibition,
any real difference between condition and concession is imperceptible.
Consequently, we will treat the conditional view and the concessive
view as one and the same.
But what about the difference between permission and condition?
Many grammarians make a distinction between these two.17 But not
all do. No less an authority than the grammar by Blass- Debrunner
lumps the permissive, concessive and conditional uses together.18
M. Barth, in his meticulous commentary on Ephesians, does the same:
he translates o]rgi<zesqe "if you are angry," labels it a "concessive
imperative," then defines what he means by saying that "a factual
permission is granted by this imperative" (italics mine).19 It may be
significant that, almost universally, those who distinguish the two opt
for the conditional nuance, arguing that permission is closer to com-
mand. C. L. Mitton is representative: "It is quite wrong to take it as a
command or even a permission to be angry. . . here the quotation
means: 'If you do get angry. ..'"20
In this context, however, one has difficulty even determining the
difference between permission and condition. This is due to the follow-
ing prohibition, mh> a[marta<nete, which somehow governs the opening
imperative. There is very little difference between "be angry, if you
16 E.g., A. T. Robertson, A Grammar of the Greek New Testament in the Light of
Historical Research (4th ed.;
usage "Concession or Condition"; cf. also J. M. Stahl, Kritisch-Historische Syntax des
Griechischen Verbums der
Klassischen Zeit (reprint ed.;
1965) 239, 362; W. D. Chamberlain, An Exegetical Grammar of the Greek New
Testament (New York: Macmillan, 1941) 86; C. Vaughan and V. E. Gideon, A Greek
Grammar of the New Testament (Nashville: Broadman, 1979) 107; B. L. Mandilaras,
The Verb in the Greek
Non-Literary Papyri (
Sciences, 1973) §729.
17 See n. 15.
18 BDF §387. They list three uses: command, request, concession. In discussing
John 2:19 they consider lu<sate to be equal to e]a>n kai> ju<shte (which they call
concessive). And regarding our passage, they argue that it "most probably means 'you
may be angry as far as I am concerned (if you can't help it), but do not sin thereby"'--
a rendering which is normally equated with the permissive view. Cf. also H. Schlier
(Der Brief an die Epheser [Dusseldorf: Patmos, 1963] 224, n. 3) who, though calling the
imperative concessive, cites Blass-Debrunner in support.
19 M. Barth, Ephesians 4-6 (AB; 2 vols.; Garden City, NY: Doubleday, 1960) 2.513.
"L'imperatif o]rgi<zesqe en soi n'etait pas un ordre, mais une concession. . . : si vous vous
emportez, ne pechez point" ("The imperative o]rgi<zesqe is not in fact a command, but
a concession. . . : "If you are angry, do not sin").
20 Mitton, Ephesian, 168.
358 CRISWELL THEOLOGICAL REVIEW
must, but don't sin" and "if you are angry, don't sin." Nevertheless,
the semantic situation found in Eph 4:26 (viz., imperative + kai> + a
second verb) fits the pattern required for a conditional imperative,
though it is quite rare for permissive imperatives.21 Consequently, we
will treat permissive and conditional as one--and, out of deference to
conditional advocates, call this approach simply the conditional view.
To sum up: the live options in Eph 4:26 are only two: either
o]rgi<zesqe is a command or a condition. We now need to examine
several factors which may help us to come down from the fence on
one side or the other.
II. Factors Contributing to the Use of o]rgi<zesqe in Eph 4:26
There are four major factors which help shape our understanding
of the nuance of o]rgi<zesqe in this text: (1) the use Paul makes of
Ps 4:4; (2) the context; (3) the general biblical teaching on man's
anger; and (4) the specifics of the syntax of the construction. For
reasons which should soon become obvious, we will treat the first two
in this section and treat the syntax separately. However, as our purpose
is to see what contribution Eph 4:26 makes toward the biblical teach-
ing on human anger, and not vice versa, we can only touch on this
third category in our examination of the context.
A. Paul's Use of Ps 4:4
As we mentioned earlier, Paul quotes verbatim the LXX rendering
of Ps 4:4: o]rgi<zesqe kai> mh> a[marta<nete. There are problems with this
translation, however. o]rgi<zesqe renders vzgr which though an impera-
tive, might not mean "be angry." The basic significance of the stem,
zgr, is simply "tremble, shake,"22 which may involve--in a given
context--shaking out of fear, trembling in awe or reverence, or shak-
ing in anger. Though the LXX renders vzgr, as "be angry," the Targum
as well as
mentaries are divided on the issue,24 though those who affirm the
21 But cf. John 19:6 (la<bete. . . kai> staurw<sate) and Rev 22:11 (a]dikhsa<tw
. . . kai> . . . r[upanqh<tw); yet even here these "permissive" imperatives bear the sense of
reluctance or toleration rather than positive permission.
22 A. Bowling in Theological Wordbook of the Old Testament (ed. R. L. Harris,
L. Archer, and B. K. Waltke [
meaning of this root is to quake or shake, from which ideas such as shaking in anger,
fear, or anticipation are derived." Cf. also BDB, 919; KB, 872.
23 NvFHt xlv hynm vfz ("tremble [in fear] and you will not sin"); klonei?sqe.
24 In defense of "be angry," cf. F. Delitzsch, Psalms (vol. 5 in Commentary on the
Testament.by C. F. Keil and
F. Delitzsch [reprint ed.;
Wallace: ]ORGIZESQE IN EPHESIANS 4:26 359
LXX's rendering tend to do so precisely because Paul quoted it.
Whether this Psalm is to be connected with the previous one,25 and if
so, whether v 4 is addressed to Absalom's men26 or David's com-
panions,27 are questions difficult to answer. My tentative preference is
to opt for the meaning "tremble (in awe)," for vzgr because (1) the
nuance of anger is rare for zgr, and is perhaps never found in the Qal
stem;28 and (2) the parallel with the rest of v 4 ("meditate. . . and be
still") seems to be a fitting balance with the idea of "tremble (in awe)
and do not sin,"29 But even if "be angry" is the meaning of vzgr,
because of the question mark over who is being addressed as well as
the object of the anger, we cannot be dogmatic about the force of the
All of this, however, is a moot point, Paul does not here use one
of his standard introductory formulas;31 he is not putting his apostolic
stamp of approval on the LXX's rendering. In my judgment, Abbott's
dictum is correct: "It is . . . superfluous, as far as the present passage is
concerned, to inquire what the meaning of the
not arguing from the words, but adopting them as well known, and as
expressing the precept he wishes to inculcate."32 His use of the Psalm
therefore, rhetorical. Hence, we need to look at the context into
which Eph 4:26 is set for further clues on the use of o]rgi<zesqe.
1976]) 114-15; W. Kay, The Psalms (London: Rivingtons, 1871) 14; in defense of
"tremble (in awe)," cf. C. A. Briggs, Psalms (ICC; Edinburgh: T. & T. Clark, 1907) 34;
H. C. Leupold, Exposition of the Psalms (Columbus, OH: Wartburg, 1959) 68-69.
25 So Delitzsch, Psalms; differently, Briggs, Psalms.
26 So Delitzsch, Psalms.
27 So Kay, Psalms.
28 "be angry, AV. is sustained by Is. 2821 of God's anger and Pr. 299 of man's. But
in these cases it is rather the quivering and trembling of passion, which is justifiable;
and is regarded by many as Hiph v. BDB" (Briggs, Psalms, 34).
29 Not only are the imperatives taken naturally as commands, but "tremble (in
awe)" and "meditate" are both
God-ward actions. The
angry your hearts, do not do wrong; though you lie abed resentful, do not break silence")
seems a bit forced.
30 Even if addressed to Absalom and his men, the idea may well be "be angry (at
your own wrong-doing) and stop sinning." Yet, if these are the addressees, Paul's use of it
is decidedly rhetorical, for he is addressing the community of believers. In large
measure, the use of vzgr just like o]rgi<zesqe, is a problem of syntax (see section III for
discussion of both).
31 In Ephesians, however, he uses an IF only twice (4:8; 5:14). See J. P. Sampley,
"Scripture and Tradition in the Community as Seen in Ephesians 4:25ff," ST 2 (1972)
101-9, for an interesting view on Paul's use of theOT in this section.
32 Abbott, 139-40. Cf. also MeyerK, Ephesians, 252; Lange, Ephesians, 169. This is
not to say that the quote has no significance, for the very familiarity of the Psalm (at
least to Paul) renders the two indicative views (discussed earlier) as highly unlikely.
360 CRISWELL THEOLOGICAL REVIEW
B. The Context of Eph 4:26
There are at least seven contextual factors, of varying weight,
which may be helpful in shaping our understanding of this elusive
1. Parenetic Section. As we have already mentioned, Eph 4:25-
32 is a specifically parenetic section in this epistle. On a mechanical
level, this might tend to favor the command view, for the other ten
imperatives here must all be taken as commands (or prohibitions).33
At the same time, none of the imperatives--except o]rgi<zesqe --fits the
structural requirements for a conditional imperative (viz., impera-
tive + kai> + second verb), which might indicate that a conditional
imperative was on the apostle's mind.
2. Community of Faith. Not only is v 26 in a parenetic section,
but it is in one which addresses the relationship of individual believer
to individual believer. It begins and ends with two indicatives ("we
are [e]smen] members of one another" in v 25; "God in Christ has
forgiven [e]xari<sato] you" in v 32), which speak of the divine initiative
toward those who now constitute the believing community. All this is
to say that, however we take o]rgi<zesqe, it should be seen as anger
directed within the church. By extension, perhaps, it can apply to
those outside the faith, but I doubt if that is the apostle's primary
point. Consequently, those who argue for the command view on the
basis of a righteous indignation toward unbelievers have missed the
thrust of the apostle here.34 But this cuts both ways: if Paul is not here
speaking about judging the world per se, then arguments against the
command view which presuppose that he is are equally invalid.35
3. A Specific Situation in View? Not to be discounted entirely
is the possibility that Paul has in mind a specific situation in 4:25-32.
Formally, all the injunctions are directed toward the group except
one. o[ kle<ptwn (v 28) may well refer to a specific individual. Not only
is it singular, but the negative mhke<ti ("no longer") indicates that the
33 This argument is helpful against seeing o]rgi<zesqe as an indicative here, but
probably not against taking it as a specific type of imperative.
34 So E. K. Simpson (Ephesians [NICNT; Grand Rapids: Eerdmans, 1957]) gen-
eralizes the passage so as to include individual nations as well as the world (108-9); cf.
also C. R. Erdman, The Epistle of Paul to the Ephesians (
35 So Yeager, New Testament, 308: "Why should he allow his anger to persist until
it has him shouting at the poor defenseless slaves of Satan who cannot help behaving as
they do since they are only unregenerated human flesh?" J. L. Boyer, too, seems to hold
this view ("A Classification of Imperatives: A Statistical Study," GTJ 8  39).
Wallace: ]ORGIZESQE IN EPHESIANS 4:26 361
stealing was already taking place.36 If this exegesis is valid, then the
entire pericope might center on this problem, and the injunction in
v 26 would then probably mean "be angry about the fact of such sin
in your midst and do something about it'"37
However, identifying a specific problem in this epistle is noto-
riously difficult. It depends not only on taking o[ kle<ptwn as referring
to an individual (rather than generically), but on seeing other specific
problems addressed in this epistle,38 as well as viewing Ephesians as
primarily intended for one church.39 More than "one thief" will be
It required to overturn the well-worn view of the epistle as some sort of
Consequently, this parenetic section is probably very loosely
organized. The rapid-fire imperatives march on asyndetically;41 these
staccato exhortations are typically Pauline.42 But even this tends to
support the command imperative view, though hardly conclusively.43
36 The substantival participle also implies this.
37 On the assumption that the thief had not yet been identified, the pericope might
have the following force:
v 25: each man should be open and honest with his neighbor--and not suspect
everyone in the community of stealing--because we are members of each other.
vv 26-27: either "be angry" at the fact of such sin within the community of
believers (cf. I Cor 5:1-5) and resolve to do something about it quickly; or, less likely,
“if you are angry” stop sinning by allowing your anger to be vented on everyone you
v 28: rebuke of the thief directly, which fits in well with the command imperative
view (at least for Paul; again cf. I Cor 5:1-5).
vv 29-31: rebuke of the congregation: the rest of you have sinned, too. As the thief
has robbed you physically, you have robbed yourselves spiritually (note the interchange
between xrei?a in v 28 and v 29)--by suspicious innuendo (v 29) and an escalating
vituperation (v 31)--which grieves the Holy Spirit (v 30).
v 32: Because of this one thief in your midst, you have forgotten Christian graces."
But, remembering what God in Christ has done for you, forgive one another.
38 But cf. C. Rogers, Jr., who makes a plausible argument for the problem of
drunkenness due to the Dionysian cult in 5:18 ("The Dionysian Background of Ephe-
sians 5:18," BSac 136  249-57); nevertheless, the Dionysian cult was not a problem
39 Even if e]n ]Efe<s& in 1:1 is original, this does not, of course, mean that the letter
was not intended to be circular.
40 Cf. D. Guthrie, New Testament Introduction (
1970) 515. Moving even further away from a specific destination/specific occasion
view, W. G. Kummel, citing J. N. Sanders, argues that Ephesians may well be "the
spiritual testament of Paul to the church" (Introduction to the New Testament rev. ed.;
Note vv 26, 28, 29, 31.
42 Cf. I Thess 5:15-22; Rom 12:9-17.
43 See n. 32.
362 CRISWELL THEOLOGICAL REVIEW
4. Eph 5:1. Barth brings 5:1 (''as beloved children, be imitators
of God") into the discussion: "Among the saints who are 'God's
imitators' (5:1) such anger cannot be excluded any more than in God
himself. . . or in the Messiah (Mark 3:5, etc.)."44 This, too, would tend
to support the command imperative view45--if all the moral attributes
of God are to be copied by the believer. But, at best, this is only an
5. The Audience. One factor rarely considered is how the audi-
ence would have understood Paul's words. Assuming that it was
largely Gentile, it may be significant that, among the Greek philoso-
phers, only the Stoics categorically condemned human anger.46 Though
the general tenor among the Greeks was a negative assessment, "the
moral wrath which protects against evil"47 was seen as entirely legiti-
mate in the realm of government and "even necessary for great acts
and virtues. . . “48
With reference to the Jewish contingency among Paul's addres-
sees, both the OT49 and rabbinic literature50 considered righteous
human indignation to be legitimate.51 On the other hand, Philo had a
difficult time accepting either human wrath or divine wrath as a
righteous emotion/ act.52 This, of course, is in keeping with his Stoic
In other words, few Jews or Gentiles in the first Christian century
would flinch at reading o]rgi<zesqe as a command. In the least, since
the Stoics and Philo stand apart from the rest of the ancient world,
those exegetes who would absolutely prohibit human anger53 might
do well to take stock of the company they keep! Nevertheless, what
the original audience would think is not conclusive for what an author
44 Barth, Ephesians, 2.513.
45 However, Barth himself sees o]rgi<zesqe as permissive (=conditional).
46 H. Kleinknecht" "o]rgh<," TDNT 5.384-85.
47 Ibid., 384.
49 Cf. Exod 32:19; Judg 9:3; 1 Sam 11:6; 2 Sam 12:5; Neh 5:6. J. Fichtner points out
that "Saul's wrath against the Ammonites. . . . is attributed to the Spirit of Yahweh (1 S.
11:6)" (ibid., 394), and further that ". . . one can speak esp. of holy and righteous anger
when it is a matter of directly championing the cause of Yahweh. . ." (ibid.; see
50 See references in Str-B 3.602 (on Eph 4:26).
51 The OT, however, seems to view it, at times, as a virtue, while the rabbinic
material simply allows for it.
52 TDNT 5.417.
53 Boyer, "Conditions," though he advocates the conditional view, categorically
prohibits anger to men: "it seems impossible to understand this in a good sense. . . .
'righteous indignation' seems never to be approved for men" (39).
Wallace: ]ORGIZESQE IN EPHESIANS 4:26 363
meant. Two final (and related) contextual arguments are usually
judged as decisive clues to Paul's meaning here.
6. ]Orgh< Prohibited in v 31. What is normally perceived to be
the strongest argument54 against taking o]rgi<zesqe as a command is the
prohibition against anger In v 31: Let all bitterness and wrath and
anger [o]rgh<] and clamor and slander be put away from you, with all
malice."55 Formally, it is not just o]rgh< that is prohibited--but pa?sa
o]rgh< (“all anger").
Vv 26 and 31 clear stand in tension. Just as it would be wrong-
by appealing only to 26a--to say that all anger is a righteous duty laid
on the believer at all times, so too it would be wrong--by appealing
exclusively to v 31--to say that all anger is wrong and utterly sinful at
all times. Indeed, there are two internal clues which help to resolve
the tension created by v 31.
First, as many commentators point out,56 this verse apparently
gives a progressively climactic and inherently cohesive list of vices;
hence, the o]rgh< which springs from qumo<j (which, in turn, is rooted in
pikri<a) is to be shunned at all times. As C. Hodge points out, “Verse
31 is not inconsistent with this interpretation [viz., that there is a
righteous anger], for there the context shows [that] the apostle speaks
of malicious anger--just as ‘all hatred’ means all malice, and not the
hatred of evil."57
Second, the very fact that Paul distinguishes between anger and
sin in v 26 indicates that there is an anger which is not sinful. Now it
might be objected that this is begging the question because it pre-
supposes an injunctive flavor for o]rgi<zesqe. But that is not the case.
Even if we assume the conditional view, “if you are angry, then do
not sin" at least implies that it is possible to be angry without sinning.
As A. Tholuck has aptly remarked, “Spricht Paulus von einem verwerfli-
chen Zorne, wie kann er das Sundigen vom Zurnen trennen?"58 And
once it is recognized that the apostle admits of a non-sinful anger ill
54 So Gnilka, Epheserbrief, 235.
55 RSV translation.
56 Eadie, Ephesians, 348-49; B. F. Westcott, St. Pauls Epistle to the Ephesians
(London: Macmillan, 1906) 74; Barth, Ephesians, 2.521.
57 Hodge, Ephesians, 270.
58 “If Paul speaks [only] of a reprehensible anger, how can he distinguish between
sinning and being angry?" A. Tholuck, Philologisch-theologische Auslegung der Berg-
predigt Christi nach Matthaus (Hamburg: Friedrich Perthes, 1833) 186. The underlying
difficulty for the conditional view, in this regard, is that it cannot handle the apodosis,
mh> a[marta<nete. To maintain both a conditional o]rgi<zesqe and an absolute prohibition
of anger requires a declaration in the apodosis, not a prohibition: "If you are angry, you
364 CRISWELL THEOLOGICAL REVIEW
v 26, then it must be conceded that he does not absolutely prohibit
anger in v 31. Therefore, "conditionalists" who appeal to v 31 prove
too much: they undercut their own view of o]rgi<zesqe in the process.
7. Eph 4:26b-27. Finally, conditionalists appeal to vv 26b-27 as
an argument against the command view. Boyer asks, " . . . if this is a
command to show 'righteous indignation,' why is the warning added
to end it before the sun goes down?"59 In response, four things can
First, if, as Boyer believes, Paul is condemning all human anger,
why would he allow it to last until sundown? Would it not be more to
the point for him to have said, "Do not get angry in the first place"?
By setting a temporal limit60 the apostle lays down a restriction, but
not a prohibition.
Second, no one who maintains the "command" view would see
o]rgi<zesqe as an unqualified exhortation. Unless it is impossible for a
command to have a limited and occasional application, it is difficult to
see the validity of Boyer's point. If I am commanded to "weep with
those who weep," is this not a limited command? Or if parents are
told not to spare the rod for that would spoil the child, does this mean
that all discipline must be corporeal--or worse, that the only way
they are to relate to their offspring is with a whip in hand? Surely the
imperative is flexible enough, in a given context, to make demands
which are limited by time and/or occasion.
Third, no one but the Stoics and Philo would deny God the
right--even the obligation--to be angry at times. Yet few would say
that anger is God's leading attribute. Isa 28:21 speaks of the exercise
of God's wrath as his "strange"61 or "unusual"62 work. The point is
that a command to be angry-and yet to limit that anger--is in
keeping with God's character and may well be, as Barth has noted, a
specific application of Eph 5:1: "become imitators of God," Does not
the psalmist say, "His anger is but for a moment, His favor is for a
Finally, entirely apart from these considerations is the possibility
that we have misconstrued the limitation in v 26b, Paul might not be
placing a temporal limit on one's anger. When he says "do not let the
sun go down on your anger," he does not use the obvious cognate,
59 Boyer, "Imperatives," 39.
60 Which is more than likely not literal, the point being that one ought not to allow
anger to fester so as to become sin. Even righteous anger, then, can degenerate, if not
61 So NEB.
62 So NASB.
63 Ps 30:5 (NASV).
Wallace: ]ORGIZESQE IN EPHESIANS 4:26 365
o]rgh<. Instead, he uses paraorgismo<j. This is a rare term which has
been found to date only in biblical Greek.64 “In the LXX it is used as a
rule with an active meaning. . . .”65 In fact, we may go so far as to say
that the term always has an active meaning except for one variant
found in codex Alexandrinus.66 It may thus be translated “the cause of
provocation,” and always refers67 to the external cause by one party
parorgismo<j is used but once in the NT, in Eph 4:26. Perhaps commen-
tators are too hasty to label it a passive--viz., the feeling of being
provoked.68 If it bears its normal
sense of “that which caused
cation” Paul might well be saying, “Deal with the cause of your anger
immediately.” And if that cause is another brother (as would be most
natural in this section), the point might well be the same as Matt
18:15: “if your brother sins, go and rebuke him.” V 27 then would
have the force of --don't let the devil gain a foothold in the assembly
by letting sin go unchecked."69 Further, mh> a[marta<nete in this view
would have the force of “do not sin by doing nothing--act quickly to
discipline your brother.” If this reconstruction is correct, then o]rgi<-
zesqe would have to be taken as a command.70
Perhaps we are reading too much into the text in this approach.
But suffice it to say here that, whether 26b is a temporal limit on one's
anger or whether it is an incitement to carry out church discipline
quickly, there is no good reason to object to o]rgi<zesqe as a command.
To sum up the contextual arguments: none of the seven points we
have made is decisive. At this stage, o]rgi<zesqe could be either a
64 As well as in patristic comments on Eph 4:26. My perusal of Thesaurus Lingua
Graece (via the Ibycus computer-generated concordance) turned up no new instances.
65 J. H. Moulton and G. Milligan, The Vocabulary of the Greek Testament (reprint
66 Cf. 3 Kgdms 15:30; 4 Kgdms 19:3; 23:26; Neh 9:18, 26. The variant reading is
found in Jer 21:5. As well, the close cognate, paro<rgisma is also found only with an
active meaning (3 Kgdms 16:33; 20:22; 2 Chron 35:19).
67 The v:l. excepted.
68 For an active sense, cf. H. C. G. Moule, Studies in Ephesians (reprint ed.;
69 See n. 37 for a possible reconstruction of the incident, if any, that the apostle
might have had in mind.
70 What might give further support for this view is the fact that o]rgi<zw, rather
than qumo<w, is used. If a distinction can be made between these two--though, admit-
tedly, there is a great deal of overlap—o]rgi<zw tends to accent the volition, while qumo<w
tends to stress the emotion (though it is probably impossible to extricate emotions
entirely from o]rgi<zw's connotati~ns). If such a volitional emphasis, is on the apostle's
mind (a nuance difficult for English-speaking natives to grasp for anger, be angry),
then the link with decisive action, justice, (informal) church discipline is thereby
366 CRISWELL THEOLOGICAL REVIEW
command or a condition, though I am inclined to think that the
command view has the edge.
III. The Syntax in Eph 4:26
The final factor deals specifically with the syntax of o]rgi<zesqe kai>
mh> a[marta<nete. There are three arguments to consider here, though
the first two are of minor importance.
The aspectual forces of the imperative are often treated in relation
to present time. Thus, the aorist imperative is usually considered to
mean "start to do X," while the present imperative bears the sense,
"continue doing X." The aorist prohibition71 has the force "don't start
to do X," while the present prohibition means "stop doing X;"72 If this
meaning were pressed in Eph 4:26, the idea might be, "keep on being
angry, but stop sinning."
But recent studies have shown that this way of viewing the
imperatives is quite incorrect,73 for the time element is entirely inci-
dental to the tense used and is to be derived from the context. As
K. L. McKay points out, "In the imperative the essential difference
between the aorist and the imperfective [i.e., present] is that the
former urges an activity as whole action and the latter urges it as
ongoing process."74 Consequently, this is a moot point for our present
B. The Connective kai>
Several commentators who favor taking o]rgi<zesqe as a command
make much of the conjunction joining the two imperatives. Meyer is
representative: " . . . the mere kai> is only logically correct when both
imperatives are, thought of in the same sense, not the former as
permitting and the latter as enjoining, in which case the combination
becomes exceptive ('only, however'), which would be expressed by
a]lla>, plh>n, or mo<non . . ."75 This is not a very strong argument for
71 In the NT, all aorist prohibitions in the second person employ the subjunctive
rather than the imperative.
72 Cf. H. E. Dana and J. R. Mantey, A Manual Grammar of the Greek New
Testament (Toronto: Macmillan, 1927) 299-303; J. A. Brooks and C. L. Winberry,
Syntax of New Testament
73 Cf. K. L. McKay, "Aspect in ImperativaI Constructions in New Testament
Greek," NovT 27 (1985) 201-26; Boyer, "Imperatives," 35-54: ,
74 McKay, "Aspect," 206-7.
75 Meyer, 2.53-54.
Wallace: ]ORGIZESQE IN EPHESIANS 4:26 367
the simple reason that kai> is not here connecting two naked impera-
tives, but an imperative on the one side with mh> plus the imperative
on the other. The negative disrupts any simple connection and, in
fact, probably lends a mildly adversative force to kai>: "be angry, and
yet do not sin." Still, the presence of kai> cannot be construed as an
argument against the command view and, in all probability, leans
toward it. Nevertheless, neither of these first two grammatical argu-
ments is very decisive.
C. The Semantic Situation of Conditional Imperatives
The final syntactical argument, however, may well be decisive.
Those who hold that o]rgi<zesqe is a conditional imperative must
reckon with the fact that it is followed by another imperative. This
would seem unnatural, as we might expect a future indicative--thus,
in John 2:19 we read, "Destroy this temple and in three days I will
raise it up [lu<sate . . . e]gerw?]." In Boyer's exhaustive study on im-
peratives in the NT, in fact, he states the following:
Probably the strangest and most controversial category of impera-
tives is that which seems to express some conditional element. Here it is
necessary to distinguish two groups. The first is neither strange nor
controversial; it includes a large number of instances (about 20) where
an imperative is followed by kai> and a future indicative verb [italics
mine]. . . .
The second group consists of a few passages where condition has
been proposed to explain a difficult passage.76
Boyer then lists only three passages77 which belong to this questionable
category. Eph 4:26 gets the greatest amount of coverage--and here
Boyer comes out strongly, on contextual and theological grounds, for
a conditional o]rgi<zesqe. The point is that one of the leading advocates
of the conditional view--and the only one to categorize every im-
perative in the NT--was unable to find any other conditional impera-
tive which was followed by kai> and another imperative. Boyer has
clearly felt the force of this syntactical argument and has found that
his only recourse is to argue on the basis of other factors.
But, to be sure, there are grammarians who argue that a condi-
tional imperative can be followed by another imperative. A. T.
Robertson has made perhaps the most cogent statement along these
76 Boyer, "Imperatives," 39.
77 Strangely, he includes John 2:19 in his dubious list, as well as 2 Cor 12:16 and
368 CRISWELL THEOLOGICAL REVIEW
Sometimes two imperatives are connected by kai> when the first suggests
concession. Thus Eph. 4:26, o]rgi<zesqe kai> mh> a[marta<nete. So also
e]rau<nhson kai> i@de (Jo. 7:52). Cf. e@rxou kai> i@de (Jo. 1:46). This seems
simple enough [italics mine].78
Robertson thus gives two examples (besides Eph 4:26) of a condi-
tional imperative followed by another imperative. But what "seems
simple enough" to Robertson does not help the cause of a conditional
imperative in Eph 4:26 for three reasons.
First, Robertson's identification of e]rau<nhson in John 7:52 and
e@rxou in John 1:46 as conditional imperatives is highly debatable, for
even Boyer--who would like to find such a tidy semantic parallel to
Eph 4:26--is unable to admit that any construction other than impera-
tive + kai> + future indicative involves a conditional imperative.79
Second, even if we assumed that Robertson's proof-texts were
valid, a proper parallel has not been drawn for us. In John 1:46 we
read e@rxou kai> i@de ("come and see"). This is Philip's response to
Nathanael's challenge, "Can any good thing come
If we see e@rxou as a conditional imperative, rather than entreaty, then
Philip's response means, "If you come, you will see." In John 7:52, the
Pharisees suspect that Nicodemus has become a disciple of Jesus.
ask, "Are you from
kai> i@de] that no prophet comes
Again, if e]rau<nhson is conditional, the Pharisees' retort means, "If you
search, you will see." In other words, in both of Robertson's proof-
texts the second imperative functions semantically as a future indica-
tive.80 If we applied that principle to Eph 4:26 we would get "If you
are angry, you will not sin"!
Third, there is an additional problem with Robertson's proof-
texts. The very fact that there is some doubt concerning the label of
78 Robertson, Grammar, 949. Cf. also A. Buttmann, A Grammar of the New
Testament Greek (Andover: Warren F. Draper, 1873) 290.
79 Boyer considers e]rau<nhson kai> i@de (John 7:52) to be ambiguous semantically,
fitting either the "command" or "condition" category; he does not deal with John 1:46.
80 It should be noted here that conditionalists who base their view on a supposed
conditional imperative in Ps 4:4 must also reckon with the fact that the same semantics
are operative in Hebrew. Abbott (Ephesians, 140) aptly points out:
The phrase is frequently explained by reference to what is called the Hebrew idiom (which is by
no means peculiarly Hebrew) of combining two imperatives, so that the former expresses the
condition, the latter the result, as in Amos v. 4, "Seek Me and live." But this would make the
words mean, "Be angry, and so ye shall not sin."
As well, in all 17 examples listed in GKC of this idiom, none broke away from the
"condition-consequence" idea (§110.2.(a)). Indeed, they noted that "In this case the first
imperative contains, as a rule, a condition, while the second declares the consequence
which the fulfilment of the condition will involve" (italics mine). (See also n. 22.)
Wallace: ]ORGIZESQE IN EPHESIANS 4:26 369
conditional imperative for e@rxou and e]rau<nhson is not because others
would give them a radically different nuance. The idea of injunction
or condition in these two texts is not very far apart at all. But this is
not due to a blurring in the distinction between the categories (or.
more accurately, nuances) of command and condition-otherwise
exegetes would not spill so much ink over Eph 4:26.81 Rather, it is due
to the fact that these conditional imperatives have not lost their
natural injunctive force. And it is probable that this is due to their
being linked by kai> with another imperative. We might even para-
phrase John 1:46 as "If you come--and I urge you to--you will see"
and John 7:52 as "If you search--as well you should--you will see." If
this were applied to Eph 4:26, it would mean, "If you are angry--and
you should be"!
Perhaps we are being unfair to Robertson, however. After all, he
only supplied two proof-texts, implying that there may be others.
Because of this possibility it is necessary to examine every impera-
tive + kai> + imperative construction in the NT. Altogether, there are
187 imperative + kai> + imperative constructions in the NT.82 This cer-
tainly seems like a large enough data base from which to draw some
fairly firm conclusions. I examined each one to determine whether we
can add any more potentially conditional imperatives to Robertson's
list. The answer is a qualified yes. In addition to John 1:46 and 7:52, 21
more imperatives can be added to the list83 of potential conditional
imperatives. I broke these down into two groups: those which only
had a slight chance of deserving the label and those which, in their
contexts, looked like good candidates. In the first group belonged 17
imperatives.84 For example, Mark 2:9 has "rise and take up your bed
and walk." It is just possible that the force is, "If you rise and take up
your bed, you will walk." Yet, the whole tenor of the pericope
81 Cf. also John 2:19 where such a blurring of nuances would wreak exegetical
82 These data were derived from Gramcord. Gramcord is a copyrighted software
package which is able to perform grammatical searches in the Greek NT. It is distri-
buted solely by the Gramcord Institute, 2065 Half Day Road,
By creating a contextfield of twelve words, 289 imperative + kai> + imperative
constructions were found. We made it this broad in order to pick up every legitimate
construction. Gramcord, however, did not discern whether such imperatives belonged
in the same clause; as well, it multiplied the examples when more than two imperatives
were used (e.g., Mark 2:9 [which reads e@geire kai> a#ron . . . kai> peripa<tei] was listed
four times). Consequently, this list of raw data was pared down to 187 legitimate
83 Eph 4:26 being omitted from consideration as that is our target passage.
84 Cf. Matt 9:5; 11:29; 15:10; Mark 2:9; 5:19; 7:14; 9:50; Luke 5:4, 23; 24:39; John
4:35; 5:8; 20:27; 1 Cor 11:28; 15:34; Gal 5:1; Eph 5:14.
370 CRISWELL THEOLOGICAL REVIEW
seemed to render this unlikely.85 In Luke 24:39 Jesus, in his resur-
rected body, says "Touch me and see. . ."; the force could possibly
be “If you touch me, you will see." But again, the tone of the passage
seems to be against this.86
In the second group--the likely candidates--belonged, besides
John 1:46 and 7:52, only two other texts. In Luke 7:7 we read of the
centurion's request that Jesus heal his servant: "Say the word and let
my servant be healed." Many scribes changed i]aqh<tw ("let him be
healed") to i]aqh<setai ("he will be healed"),87 indicating that the
second imperative is virtually the equivalent of a future indicative. "If
you say the word, he will be healed," is not an inappropriate render-
ing, therefore.88 John 11:34 reproduces the verbage of 1:46 ("come
and see") and consequently may well imply a conditional nuance.
Significantly, of all 21 potential example, only two were as con-
vincing as Robertson's two alleged proof-texts. Thus, out of 187
imperative + kai> + imperative constructions in the NT, four proba-
bly--or, at least, quite possibly--involve conditional imperatives. Yet
each of these four could be construed as conditional imperatives
precisely because the trailing imperative functioned as a future indi-
cative--a semantic situation which finds no parallel in Eph 4:26.
However, among the 17 mildly possible conditional imperatives,
I found a different phenomenon. In four passages,89 assuming that the
first imperative was conditional, the second still, most naturally, bore
its injunctive force, thus paralleling Eph 4:26. However, there were
two major problems with all these examples: first, they were exceed-
ingly doubtful as legitimate candidates for conditional imperatives;
and second, the conditional imperative nuance still carried with it the
full force of a command. Two examples should suffice. In Mark 5:19
Jesus told the formerly demon-possessed man, "Go home and tell
them what the Lord has done for you." If we read u!page conditionally
85 soi> le<gw in v 11 sounds like it introduces a command; the man's immediate
response suggests that he viewed it as a command; and the fact that Jesus stresses his
own authority (v 19) would best fit a command imperative. See also Matt 9:5; Luke
5:2.3; and John 5:8 for the same expression.
86 The parallel in v 39a and the apparent eagerness of Jesus to get his disciples to
believe in him are decidedly on the side of seeing entreaty/command here.
87 In fact, only p75 B L 1241 copsa,bo are listed in UBSGNT3 in support of the
imperative. A quick check of The New Testament in Greek: The Gospel According to
St. Luke, Part One: Chapters 1-12 (IGNTP; Oxford: Clarendon, 1984) revealed no
88 Cf. also the v.I. in Gal 6:2 (a]naplhrw<sate) where the UBSGNT3(=NA26) text
has the future indicative.
89 Matt 15:10; Mark 5:19; 7:14; Luke 5:4.
Wallace: ]ORGIZESQE IN EPHESIANS 4:26 371
we--would have, "If you go home, tell them. . ." Though a command
would thereby be preserved in the apodosis, only with great ingenuity
could we construe u!page as a mere option.90 Luke 5:4 suffers the same
judgment, for Jesus' command to Peter to "Put out into deep waters
and lower the nets" can hardly, without torture, be rendered, "If you
put out into deep waters, lower the nets." The context must virtually
be suffocated to get this idea out of the verse.91
It must be readily admitted that these examples are very difficult
to swallow. They are included in this discussion to show that only by
great mental gymnastics is one able to show legitimate parallels to a
conditional o]rgi<zesqe in Eph 4:26.
To sum up the major syntactical argument we can make the
following three points:
(1) All the positively identified conditional imperatives in the NT
are followed by kai> + future indicative.
(2) All four of the probable conditional imperatives in impera-
tive + kai> + imperative constructions require the second imperative
to function semantically as a future indicative (i.e., stating the con-
sequence/fulfillment of the implied condition).
(3) All of the 21 potentially conditional imperatives in impera-
tive + kai> + imperative constructions retained their injunctive force.
These three syntactical facts I consider to be decisive against a
conditional o]rgi<zesqe because the semantic situation of conditional
imperatives is so radically different from what we see in Eph 4:26.92
(In light of this, we might well consider the distinct possibility that
what the phenomena of the NT display is hardly unique to itself: the
semantic pattern of conditional imperatives found in the NT might
just be an aspect of universal grammar as well.) Furthermore, the
normal expediency of appealing to the use of Ps 4:4, the context, or
the general biblical teaching on human anger as that which must
override any notion of command in o]rgi<zesqe is inconclusive at best,
and, as we have hopefully shown, more than likely supports the
command view. Eph 4:26, then, can be taken at face value: "Be angry
and do not sin."
90 Jesus had just prohibited him from coming with him. This alternative, then, is
not "if you go home rather than coming with me" because the latter was already
91 In particular, Peter's response in v 5 indicates that he would have been unwilling
to do this except that Jesus commanded him.
92 If one wishes to debate whether this verse or that belongs in the category where
I have placed it, such would not invalidate these three points. We could just as easily
drop the numbers and say, . . All of the potentially conditional imperatives. . . ," etc.
372 CRISWELL THEOLOGICAL REVIEW
IV. Conclusion and Application
In Eph 4:26 Paul is placing a moral obligation on believers to be
angry as the occasion requires. As his injunction is in a parenetic
section dealing with how believers are to interact with each other-
rather than with the world--he probably has in mind a righteous
indignation which culminates in church discipline, though not neces-
sarily in a formal way. Since this righteous indignation is a part of our
response to imitate God, it must be an "enlightened wrath, the wrath
whitened by grace."93 As God himself does not dwell in anger, neither
should we. As anger is the dark side of God--his strange work--so
too wrath must never characterize the believer. However, if we fail to
obey this injunction, not only will the enemy continue to make well-
ploughed inroads into our churches, but we ourselves will, by sup-
pressing our holy indignation, be but "a maimed sample of humanity."94
93 M. B. Lang; "Isaiah 1.18 and Ephesians IV.25-29," ExpTim 8 (1896-97) 405.
94 Simpson, Ephesians, 108.
This material is cited with gracious permission from:
Please report any errors to Ted