Truth, in the Jewish tradition, belongs preeminently to God, he who “is true, whose word is truth and holds true forever.” The attribute of truth attaches, as per the quotation, to that which God says, both in terms of present veracity (God does not falsely represent the facts as they are), and in terms of prospective reliability – God will not break a promise or renege on a commitment normative or otherwise. And it not only attaches to but in some sense is identical with God himself: as Jeremiah says, “The Lord God is truth” (Jer. 10:10). God is someone you can count on, categorically and without remainder.
For the psalmist(s), this fact coupled with its implicit contrast class forms a central fulcrum of faithful religious life and practice. God, and God alone, is worthy of trust: “Happy is the man who makes the Lord his trust, who turns not to the arrogant or to followers of falsehood [kazav]” (Ps. 40:5). Why? Because whereas God is “my deliverance and glory, my rock of strength, my refuge,” mere mortal men are “but breath” and “falsehood” (Ps. 62:8, 10). The dichotomy seems clean and complete – God is true, people are not – though intriguingly, in one place the psalmist betrays a certain self-awareness as to the possibly overzealous, situationally provoked character of the unqualified generalization: “Out of great suffering I spoke, and said rashly, ‘All men are false’” (Ps. 116:11). Collating the data, we find that (a) God alone is categorically true; (b) humans may prove true, and deserve scorn when they don’t; but (c) humans are on the whole paradigmatically unreliable, with falsehood close to their core.
Were it only that we could reserve our faith for the Lord alone, but the fact is that in life we have little recourse from at times placing our trust, however partially and provisionally, in our fellow, falsehood-riven mortals as well. We require, therefore, a way of discriminating potential objects of trust more nuanced than the either/or of truth vs. falseness, emet or kazav. Leaving eternal judgments aside, in the course of life we need to be able to determine whether an individual or class of persons with this or that record of truthful or deceptive comportment ought to be trusted for such and such a purpose under some or another set of conditions. Can the Jewish tradition help with that?
It is important that in talmudic thinking lying is not, on the prevailing consensus, categorically forbidden. “How ought one to celebrate [dance] before a bride,” it is asked? According to Beit Shammai, strict honesty must prevail: “[according to] the bride as she is” (Ketubot 16b). That is, nothing should be said or indicated of the bride that exceeds the bare truth of her excellences. She is as she is, and not everyone is a winner. But Beit Hillel maintains that, regardless of the reality at hand, one ought to expressively recognize her superlative qualities – “A beautiful and gracious bride.” Surely one mustn’t speak the truth when that truth is hurtful or upsetting to no end, and so if in such a situation one must nonetheless speak, surely it is some measure of falsehood that ought to be spoken. Not only, therefore, is there no categorical prohibition on lying, but the uttering of falsehoods can be positively enjoined: Where R. Eliezer ben R. Shimon allows that it is “permitted to adjust [the truth] for the sake of peace,” R. Natan goes the final yard, classifying such dissimulation a “mitzvah” (Yevamot 65b).
But for lies that do not fall under the peace exemption, how ought we to regard their perpetrators? For a variety of legal functions, candidates for participation are classified with respect to their ‘credibility’ [ne’emanut] as a determinant of their qualification for the role. A prospective witness meeting all the standard criteria for eligibility (e.g. an adult Jewish male), for instance, may be disqualified from testimony on account of his prior conduct, and this in various ways. Most powerfully, those known to have committed certain grave sins, such as Sabbath violation or murder, are unanimously ruled resha’im and thus ineligible out of hand. For lesser crimes there are several disputes, with the various rulings depending on the severity and form of the prior violation relative to the judicial proceedings at hand (see Sanhedrin 27a).
Importantly, credibility is not all or nothing: Someone convicted of financial crimes and thus ineligible for testimony in financial litigation may remain eligible for testimony in capital cases on account of the greater solemnity assumed for the latter; just because someone is willing to lie for nickels and dimes doesn’t mean they’ll lie when the stakes are life and death. Similarly, according to R. Nahman (as interpreted by Rava), someone suspected of sexual impropriety remains eligible for testimony in general but is disqualified for cases concerning a woman’s marital status, particularly when the testimony, if accepted, would result in the woman’s newfound marital availability. The issue, of course, is that though the man remains perfectly trustworthy in general, for this purpose his record renders him not credible.
Here the Tosafists add an intriguing consideration to the mix. Is it not the case, they ask, that someone convicted of a high-grade sin like sexual impropriety is a rasha like any other and thus ineligible as a witness wholesale? On one approach provided, the answer is that sexual violations does not qualify one as a rasha because when it comes to sexual temptation, one’s “nature seizes him” (Tosafot to Sanhedrin 26b). That is, when it comes to sex, people can lose control – they are not themselves – and thus their missteps do not speak to their fundamental character. I lost control, got carried away, made a mistake, the line would go. That’s not me.
On the same theme, R. Nahman adds that “A Nissan-thief and a Tishrei-thief are not thieves, so long as it is a sharecropper, a small amount, and finished produce” (Sanhedrin 26b). That is, a sharecropper laboring in the prime harvest months who takes a bit more than his lawful share of the season’s yield, is, despite his theft, not rendered invalid for testimony as a rasha. Why is this theft different from all other thefts? Rashi adduces two cooperating reasons: (1) presumably the sharecropper justified it to himself [moreh hetera], reasoning that his hard work and sweat entitled him to slightly more than the contractual amount, and (2) he likely assumed that the owner wouldn’t mind (Rashi, ad loc.). What’s fascinating about this is that the sharecropper is wrong – he was not so entitled, and the owner did mind – but that nonetheless his incorrect, subjectively concocted rationalization is adequate to subvert his qualification as a thief. He stole, but because he did not believe he was doing anything wrong, he cannot be termed a rasha. He is a man who stole, but he is not a thief. Abstracting the operating principle, Yad Ramah concludes that, “Whenever it is possible to say that [the perpetrator] did not intend [to violate] a prohibition, he remains qualified for testimony.” What is required is criminal intent, and importantly, the standard of proof required for conviction is that criminal intent be established beyond a reasonable doubt.
Bridging the intentional/unintentional thievery issue with our more direct concern of lying is a fascinating ruling in Bava Metzia 5a. Suppose I confront you with a demand to return the watch I’d lent you and you deny having it. If the watch is found to have been on your person at the time of your denial, your lie renders you invalid as a witness going forward, as the lie was a clear act of intentional thievery. If, however, I had lent you money – where the presumption is that you were welcome to spend it in the interim – or if you simply didn’t have the watch on hand, your denial is non-incriminating. Why? Because, the gemara says, it may be that rather than intending to do away with my money, you were merely “pushing me off,” buying time to gather the funds. Perhaps, though you fully intended to repay the debt in good time, you were worried that if you conceded the debt, your failure to pay on the spot would be damaging in some way (Rashi ad loc.). Your concern needn’t be particularly well-founded or rational; what matters is that, within your own subjective frame, your intent was not criminal.
Of course lying in this way is presumably illegal in itself, and again, it’s not obvious that it was even a rational thing to do – coming clean from the get-go may well have been a far wiser move. But confronted unawares with an accusatory situation the eventualities of which appeared outsized and potentially, cripplingly devastating, you panic and deny the charges outright so as to preserve your options going forward. In other words, without intending a criminal outcome, you elect to control the story. And however unseemly, this, the Talmud judges, is certainly non-incriminating. It is a kind of lie that does not make you a liar.
When it comes to testimony itself, there is substantial leeway allowed in various directions to avoid the conclusion that false-testimony had been rendered. By rabbinic law, witnesses testifying as a pair (a minimum of two are required in most cases) are deposed separately with the results crosschecked for consistency. How consistent is consistent? Some vagueness, it turns out, is allowed: if one witness says the murder occurred at 2pm and the other witness says the murder occurred at 3pm, the split is accounted to inaccurate timekeeping, not willful deception. (Recall that the average person of the time did not possess a particularly accurate timepiece.) Essentially, the time-specifying words are charitably interpreted as bearing a tacit penumbra of vagueness – they clearly didn’t mean exactly 2pm or 3pm. If, however, it’s 2pm vs. 4pm, that may well be a stretch too far (Mishnah Sanhedrin 5:3). The measure of consistency will also depend on the centrality of the conflicting detail in question to the case – a divergence regarding the murder weapon is more problematic than a divergence regarding the murderer’s outfit (see Sanhedrin 41a).
There are also provisions for simple forgetfulness. Suppose James lent Herbert money, a fact which Herbert concedes. When James sues Herbert to recover the funds, Herbert counters that he paid the loan back already, and did so in front of witnesses John Doe and John Schmoe [ploni u-ploni]. But when the John’s are called to testify, they say they have no recollection. Is Herbert now a confirmed, established liar [huhzak kafran]? R. Nahman thought yes. But Rava objects: “Things which aren’t pressing upon a person, one does them without attention” (Shevu’ot 41a). That is, since John and John weren’t, so far as we know, particularly invested in the goings on between James and Herbert, that the repayment in fact occurred is perfectly consistent with their not remembering the event. The same principle applies to a litigant: If I, a defendant, deny having ever passed by the pillar where you claim you lent me money, and witnesses testify that I had, in fact, once urinated aside that very pillar, my credibility remains untarnished (see Shevu’ot 34b). When the matter wasn’t of clear import at the time, not recalling does not a liar make.
Given the conventional, decidedly non-absolute nature of linguistic signification, the meaning of words and statements can vary widely across persons and contexts, with that variance depending on any number of mechanisms, including simple stipulation: “When I say ‘red,’ I mean ‘blue’” is perfectly coherent, if perverse. The rabbis thought of this one too. When a court administers an oath to a litigant, the litigant must be notified as follows: “Be aware that it is not according to your own intention that we administer this oath to you, but rather according to our intention and the intention of the court” (Shevu’ot 29b). Without that stipulation, someone could swear that they “returned the money,” where what they had really done was hide the cash in the tip of a cane temporarily handed to the plaintiff while the oath was administered (see Shevu’ot 34b) – clearly not what the court meant by “returned the money,” but a perfectly coherent, if idiosyncratic, use of language. In other words: I said what I meant, but I didn’t mean what I said.
The limits of idiosyncratic interpretation are debated. In general, when interpreting a speaker’s utterance, do we follow their words [batar pumeih azlinan] or their intentions [batar da’ateih azlinan]? The gemara records a tradition that, just as with the litigant before a court, when Moses administered the oath to abide by the Torah to the people of Israel he too stipulated that the oath was in accord not with the people’s intentions but with his and God’s. Why was this necessary? One possibility is that, were Moses, for instance, to have said “you shall observe what God says,” the people might have repeated the formula but by the term ‘God’ meant some foreign deity. On this view, unless otherwise stipulated, anyone can use any word to mean more or less whatever they want. It is argued, however, that the people may not have enjoyed complete latitude in stipulating meanings – it may be that in this example, the word ‘God’ actually admits conventionally of both meanings, and so the alternative interpretation was available even excluding purely private intentions from the calculus. On this view, idiosyncratic meaning assignment is constrained by the range of minimally conventional usages for a given word. Presumably some version of this constraint applies on the former view as well: likely no one would allow that by ‘God’ you meant ‘the barrel of tomatoes in the attic.’ In sum, arguments of the form ‘Yes I did say that, but what I meant was…’ are potentially valid, but only within reasonable limits.
Even assuming the sense of a word string is held static, the overall meaning of a statement can vary depending on what’s called illocutionary force – not what I say, but the purpose for which I say it and the act my saying it constitutes. The words ‘it’s raining’ can be employed as an affirmation, a question, a promise, and so on. One example attended to in the Talmud is the ambiguity between speaking seriously and speaking in jest. Suppose you confront me in the street, with (hidden) witnesses present, with the fact of my debt to you, and I say ‘yes,’ apparently affirming that fact. In court the next day, you cite my admission as evidence. Now, I can’t deny that I said it. I can, however, claim that though I really did say I owe you the money, I was just kidding [meshatteh hayiti bakh] (Sanhedrin 29a). Because my statement is plausibly interpretable as a joke, the inconsistency of that statement with the truth is not damaging to my credibility. Said in jest, a counterfactual statement is not a lie.
Again, though, there are limits. If I offer my admission not in the street but in court, or before witnesses I acknowledge to be serving this judicial purpose, I cannot plausibly say that I was kidding (see Shulhan Arukh, Hoshen Mishpat 81:6). Context matters: Joking around on the street is one thing, but falsely admitting a debt in court for comedic purposes is ridiculous behavior. Considering a different kind of shift in context, the gemara poses the question whether someone would jest beside a deathbed (Bava Batra 175a). The answer is no, people don’t do that. If I say I owe a dying man money in the presence of the dying man, the presumed solemnity of the situation is such that I cannot plausibly claim that I was just fooling around. So when things are flowing, low-key, and off the cuff, I was being sarcastic is a perfectly valid excuse. But when the context implies solemnity and seriousness, I am rightly held to my word.
What emerges is that to the rabbinic mind, when it comes to the character of actual human beings, honesty and mendacity are not a zero-sum, all-or-nothing affair. Human personalities are far more complex, and far richer, than that, and the light of a fundamentally virtuous character can shine on through the darkness of occasional failure. Someone who lies, unprompted and unforced, for gain; someone who lies without possible pretense of good intentions; someone who defies pervasive standards of solemnity and seriousness: someone like that, to be sure, is a liar and a thief and surely not a gentleman. But then that is an extreme case; one, as the Talmud teaches, not applicable to the preponderantly good men and women of any republic.
 Amidah for Rosh Ha-Shanah. Translation based on The Koren Rosh Hashana Mahzor, transl. Rabbi Jonathan Sacks (New Milford: Koren Publishers, 2011), 77.
 Translations taken from JPS Hebrew-English Tanakh (Philadelphia: Jewish Publication Society, 1999).
 In arguing the point, Beit Hillel appeals to what they take must be Beit Shammai’s common moral intuition: “According to your words, if one has made a bad purchase in the market, should one praise it in front of him or denigrate it? Surely, one should praise it in front of him” (Ketubot 16b). See Nedarim 27b where Beit Shammai and Beit Hillel debate the parameters of the allowance to swear falsely for the purpose of evading unjust tax burdens, with Beit Hillel taking a consistently more permissive line.
 See also Bava Metzia 23b, where scholars are reported to regularly lie, apparently without demerit, regarding matters of “tractate, bed, and guest” [masekhta, purya, ushpiza]. For helpful exposition and analysis, see Shlomo Zuckier, “The Noble Lie in Halacha,” The YU Lamdan (http://www.thelamdan.com/from-the-masechta/the-noble-lie-in-halacha/).
 See opinion of R. Yose, Sanhedrin 27a. And in the other direction: someone convicted of purely religious crimes may remain eligible for testimony in general, as a willingness to offend God does not entail a willingness to injure fellow humans (see opinion of Rava, ibid.).
 Yad Ramah ad loc. Cf. the similar formulation employed in Ramah’s gloss to Shulkhan Arukh, Hoshen Mishpat 34:4.