On Matthew's account of the temptation of Jesus in the desert, the exchanges between the tempter and Jesus are recorded for three temptations. The last one ending with Jesus sending the tempter away: Ὕπαγε, Σατανᾶ (SBLGNT).

Focusing on Jesus's reactions to the first two unrighteous proposals, found on verses 4.4 and 4.7 of Matthew, upon translation:

4 But he answered and said, It is written, Man shall not live by bread alone, but by every word that proceedeth out of the mouth of God. ... 7 Jesus said unto him, It is written again, Thou shalt not tempt the Lord thy God. (KJV Matthew 4.4,7 emphasis added)

and (randomly) listing other translations' corresponding take for the terms above in bold, we have (this is by no means a comprehensive list): ASV: said, said / YLT: said, said / DARBY: said, said / DLNT: said, said / ESV: --, said / EXB: --, answered / LEB: said, said

However, in the CLNT, and in the original, we find: CLNT: said, averred / SBLGNT: εἶπεν, ἔφη

In the original we have two different verbs, the second one, ἔφη, being the imperfect active indicative of φημί. Zodhiates [1] first lists meanings for φημί such as "generally to say, speak", but then later admits meanings like "Emphatically, meaning to affirm, assert" and lists synonyms for it as διαβεβαιόομαι "to affirm confidently", and διϊσχυρίζομαι "to assert vehemently".

With these possibilities in mind, translating εἶπεν by "said" seems pretty straightforward, but translating ἔφη also as "said", as several translations do seems to hide an important nuance of the exchange: a possible escalation on Jesus part during the exchange (which culminated in Ὕπαγε, Σατανᾶ, as noted above) as more clearly laid out by CLNT.

With the CLNT translation I get much more out of this passage. Instead of seeing/picturing (i) a cool, neutral, serial and steady dismissal of the tempter's deceptions, or even a (ii) hungry, weakening Jesus that was sweating to do the right thing, I see a (iii) perfectly righteous, increasingly bold Jesus, actually "offended" with the devil's proposals–one that did not find the devil's temptations appealing or even attractive (as preachers say they are to many of us).

Now to the question:

I'm really impressed of how frequently translators seem to overlook things like this. It took a translation with an explicit philosophy of consistency and exactness in which different words (or lemmas) in the original must be rendered differently in the target language (English) for this nuance to show up! (I'm not endorsing the CLNT, OK? I'm citing it only because it was through the CLNT that I realized the original brings different verbs: εἶπεν and ἔφη).

Based on this case, what are the arguments (not opinions!) (pro and con) of translating differently different words in the original, even if they seem to be synonymous?

I'd appreciate if answers would explicitly name the translation philosophy(ies) that are being discussed.

  • 1
    I don't know if there's a difference between φημί and λέγω/εἶπον, interesting. A couple notes: 1) the BDAG entry for φημί; 2) Modern translations apart from the most highly literal usually collapse ἀποκριθεὶς εἶπεν to 'answered' or some such, not to contrast with φημί but because it's a specific use of λέγω/εἶπον in the (Hebraic?) 'double verbs of saying' formula, somewhat avoiding your conundrum altogether; 3) In v. 10, where one might expect the peak of this ‘escalation’ the narrator reverts to λέγω.
    – Susan
    Feb 2, 2015 at 6:34
  • Thanks. From (1) most transl. could indicate the two diff. verbs being used in the original by transl. λέγω as say and φημί as affirm (thus communicating the nuance of attitude on Jesus part)—and part of the question is why stuff like this seem to be overlooked; (2a) ἀποκριθεὶς εἶπεν—well pointed! (2b) Interesting point but I doubt this could be the case given two distinct verbs are used in the Greek—perhaps with the exact purpose of ruling this out; (3) Coming back to λέγω is a clear indication of Jesus authority over the tempter, who doesn't need to emphasize his words to send him away.
    – cnaak
    Feb 2, 2015 at 12:23
  • Re. (1) and (3) - OK. (2b): You doubt what could be the case? That ἀποκριθεὶς εἶπεν is a formulaic pleonasm reasonably rendered by a single word in English? (See also λέγω III:7 in LSJ for related expressions.) All I was getting at was that doing so here ("replied" or "answered") yields a translation that distinguishes between ἀποκριθεὶς εἶπεν and φημί (see NIV, ESV, RSV, NRSV, HCSB, NET...). I really wasn't trying to be critical, though. It seems like an interesting question.
    – Susan
    Feb 2, 2015 at 13:53
  • Re. Comment character limit prevent me from being clearer. I understood your point (2) as being two-fold: (2a) collapsing of ἀποκριθεὶς εἶπεν—also noticed and understood; and (2b) εἶπεν (v.4) and ἔφη (v.7) possibly indicating a Hebraic double verb formula (my Hebrew language knowledge is very limited). My comment (2b) refers to this alone. Please, let me know if I misinterpret anything within your point (2).
    – cnaak
    Feb 2, 2015 at 14:26
  • 1
    Yup, I'm glad we clarified - thank you for your precision. The pleonasm/"double verb formula" I meant here is ἀποκριθεὶς εἶπεν (וַיַּ֥עַן וַיֹּ֣אמֶר .... or something...my Hebrew isn't so great either). Seems like in Greek usually either one of them is a participle (as here, and zillions of times elsewhere using this expression in Matthew) or else they stick a καὶ between two finite verbs.
    – Susan
    Feb 2, 2015 at 14:59

3 Answers 3


Your question is:

Based on this case, what are the arguments (not opinions!) (pro and con) of translating differently different words in the original, even if they seem to be synonymous?

Pro distinction in translation

  1. More transparency of the original text: One cannot argue the fact that using a different word in translation for each distinct word in the original makes a translation more transparent to the fact that a distinct word was used in the original. This cues an end reader not fluent in the original language that the original text has a distinction (such as in your case, where it took such a translation to even clue you in that two distinct Greek words were used), empowering the end reader to investigate the distinction themselves more fully through target language based tools they can read. (But see Con #1.)
  2. Allows subtle variation of meaning: Using a distinct word for each does allow for a subtle variation of meaning to be communicated, if the translator believes such to be the case from the original and so desires to communicate that fact. (But see Con #2.)
  3. Can add clarity to a translation: Based on both points above, translating a distinction can add clarity to a text. (But see Con #3.)

Con distinction in translation

  1. Transparency can cause confusion: Languages have distinct ways of communicating things, not always well aligned in a way that the target language can translate effectively what the source intends (idioms are examples). So exact word for word transparency is not always the best choice to make when one is trying to make the translation understood in the target language. In the case of synonymous words, not used idiomatically, would translating a distinction potentially cause confusion; see next two points related to this...
  2. Synonyms can simply be used for variety: Because two distinct, yet synonymous words are used does not necessarily mean any "subtle variation of meaning" was actually intended by the original author. So in translation, depending upon the number of words a target language has to convey similar synonymous meanings that the source language may have intended, it may be best to translate the words the same to avoid a shift in meaning that the author may not have intended, which...
  3. Can obscure the original meaning: If a more synonymous idea is intended, with variety used only for stylistic variation, then a wrong choice in word distinction could lead one to meanings not intended by the author.

The Case of Mat 4:4, 7

Here is one of countless instances where one is faced with the fact that all translation involves by necessity some interpretation. There are at least three (broad) choices one might make in translating here:

  • Use the same word for each because the Greek translator sees them as synonymous and does not want to risk communicating anything else to the English reader that may inadvertently lead them to believe some other meaning is intended. This is what many translations have done ("said ... said").
  • Use a distinct word for each because the Greek translator sees them as having a subtle variation and seeks to communicate that to the English reader ("said ... averred"). This is what you are arguing for, and the CLNT translators felt was correct.
  • Use a distinct word for each because the Greek translator wants to communicate distinction, but at the same time, carefully choose the synonym that holds a very synonymous idea. One can imagine combinations like: (a) "said ... told" (the latter term is found in HCSB, though that Bible chooses to drop out the v.4 reference by idiomatically translating the phrase there as simply "answered," as is commonly done like this comment notes); (b) "said ... stated,"; (c) "said ... spoke." All of these show distinction without leaning toward any real distinction of meaning.


So which method is correct?

Well, that depends on the interpretation of the one translating. Is it correct to hide the distinction, so that an English reader clearly understands the terms synonymously? It is if that is what the author intended. Is it correct to show the distinction, and give a shift of meaning? It is if that is what the author intended. Is it correct to show the distinction, and give as near a synonym in the target as possible? It is if that can be done in the target language and that is what the author intended.

It boils down to whether "escalation on Jesus part during the exchange" is actually intended. You believe it is, others may believe it is not. That φημί can have the idea of "Emphatically, meaning to affirm, assert," does not mean it always should have that meaning applied, as it often is used as simple synonym to mean "generally to say, speak." In English, there does seem to be enough options that translators could show distinction without implying any real shift in meaning. Perhaps most do not do that, because they have idiomatically translated the first use by combining the phrase ἀποκριθεὶς εἶπεν into "answered" and using "said" for the second.

I personally believe showing distinction, when possible, is best, especially when two synonyms are used in nearby context to one another, which does show that either for meaning or style variation, the author chose two words (but determining which, meaning or style, is going to be the burden of the one translating).

  • 1
    First, great balanced overview. One thing, in your conclusion, second paragraph, first sentence. Do you mean it boils down to the question of if escalation is intended, instead of fact? When I first read it I thought you were making a conclusive statement but then it ended oddly and the next sentence didn't fit. Bit picky maybe, sorry, just didn't want to misunderstand what your own claim on the escalation was. But as I said great answer and I agree. Even if it is the "said...told" it's better than just pretending they are the same, almost misleading the reader.
    – Joshua
    Jun 9, 2015 at 14:17
  • @JoshuaBigbee: Well, what I mean is the fact of whether escalation is intended or not is the answer to the question of what word distinction (if any at all) should be done. I'll reword that sentence.
    – ScottS
    Jun 9, 2015 at 14:37

The only remarkable observation is that this Greek word is both the imperfect and aorist active indicative (3rd Person Singular) of the verb φημί.

In this regard, Robinson (1919) cites the following (emphasis added):

If one is surprised to see this verb put under the list of second aorists, he can turn to Blass, who says that it is “at once imperfect and aorist.” It is common in the N. T. as aorist (Mt. 4:7, for instance, ἔφη). It is not always possible to decide.1

Blass (1911), whom Robinson cited, above, provides the following:

—From φημί, except for -μί, -σί, ἔφη (which is at once impf. and aor., as in Att.), no [perfect] forms are represented in N.T.2

Why then the appearance of φημί in Matt 4:7?

Perhaps there was emphasis in mind (imperfect aspect of continued action), since the quotation of the Torah by Jesus comes from the immediate context following the "Shema" (Deut 6:4). That is, the "Shema" (obeying the Lord) is the priority of interpreting Scripture, which Satan was twisting by citing Psalm 91:11-12 (and thus was suggesting that Jesus jump from the temple summit).


  1. Blass, F. (1911). Grammar of New Testament Greek (H. S. J. Thackeray, Trans.). London: Macmillan and Co., Limited, 50.
  2. Robertson, A. T. (1919). A Grammar of the Greek New Testament in the Light of Historical Research. Logos Bible Software, 310-311.

Question Restatement: Based on this case, what are the arguments (not opinions!) (pro and con) of translating differently different words in the original, even if they seem to be synonymous?

Answer: To maintain the same level of Cognitive Internalization, the words would/should be translated distinctly to convey dramatic sense--but not to the point of conferring the wrong dramatic sense.

I am not certain exactly the focus of the question:

  1. Different words should be chosen.
  2. Which word, seems like a lengthier debate, because there are other choices, than what is presented.
  3. And, how often that translation should be revisited, to pick a better translation, (which can be quite often).

Denotative Readings

When the denotative meaning of a word is used, (the strict semantic meaning is chosen):

  1. The translation does not have to change that often, as denotative meanings do not change often--but heavy connotation does change often.
  2. The modern reader loses the same/dramatic effect that the original reader would have had.
  3. Word Choices are often based on tradition, a doctrine, or to maintain a low reading level.

Cognitive Internalization

Methodologies for choosing words distinctly in translations is to try and invoke the same cognitive processes involved for the earliest readers, as modern readers.

Much of Scripture is written with Poetic and Philosophic flare, and many words have connotations that are incredibly deep--and have "Dramatic Sense".

Cognitive Internalization is a mechanism whereby we infer a pragmatic meaning of statements, preserving connotation--not just the the denotative, semantic meaning of words.

These cognitive processes include abstraction, interpolation, association, etc, and the "feeling" of a passage is attempted at, not just the "strict meaning".


This is problematic, if the translations do not change. Words heavy in Connotation frequently change in Connotation.

For example:

At one point "Heart" was a great translation for "Love with all of your heart," but where this once implied "loving with all of your thoughts, and intents", now it implies ... emotions.

For "Amaze", it was understood to be "astounded, perplexed". Now, with Spiderman, it is almost Synonymous with a "Joyful Surprise".

Issues of One-to-None Distinct Translations:

This leads to complete confusion, meaninglessness, and the injection of many doctrinal presuppositions.

Occurs when words are Transliterated, instead of Translated.

Christ from Khristos Messiah from Meshiakh Baptism from Baptidzo Graze from Kharis etc.

Issues of Many-to-One Distinct Translations:

Non-Distinct translations also have a tendency to devolve into absurdity, making distinctions where there are none, presupposing the reader is not smart enough.

These translations propose many translations, to one word in the original text.

For example:

There is one Greek work behind the words "Trust, Believe, and Faith", (pistos).

When we see "Faith" In English, our Cognitive Internalization Processes fire up, and we invoke "Martin Luther!", and "Faith not Works!", and "Blind Faith!"

That is nowhere near the Cognitive process the Greeks went through. They saw a word, "Trust", and from its context, inferred whether that trust was substantial, based on the form of the existing relationship.

Those inferences processes cause "Dramatic Effect", and cause internalization and "Crystalization" of certain concepts with considerable emotional connotations, wrongful ones.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.