Sam Harris has made his short eBook Lying (2011) available to download for free. Make the most of this opportunity and read it. It will only take you an hour or so. For those of you who cannot spare an hour, here’s a five-minute review and five-minute criticism. To the review.
In good philosophical fashion, Harris opens his essay with a definition of lying: ‘To lie is to intentionally mislead others when they expect honest communication’ (p. 4).
The expectation that others have about a particular piece of communication is important, Harris contends. When, for example, someone greets us in the street with a ‘Hi, how are you?’ an acceptable response is ‘Well!’, even when we aren’t well. Although such a response may be deceptive, Harris argues that it’s not a lie because both parties understand that the communication is intended as a greeting. By responding with ‘Well!’ you’re not intentionally misleading your interlocutor. They don’t expect an honest reply.
Our intent is also important, especially for distinguishing between communication that is true and that which is truthful, Harris argues. To speak truthfully is to accurately represent your beliefs. Those beliefs may be wrong, in which case they are false, but as long as you are not deliberately manufacturing falsehood your communication is truthful. You are not lying when you are being truthful.
With Harris’s preliminary definition and distinctions out of the way, let’s turn to his approach to ethics. Harris is a consequentialist. Consequentialists believe, unsurprisingly, that the consequences of an action determine its ethical status. If an action leads to positive consequences, it’s good; if it leads to negative consequences, it’s bad. However, actions have immediate and longer-term consequences. When it comes to telling lies, and especially to telling ‘white lies’, we tend to focus on the short-term consequences. But if we’re mindful of the longer-term, Harris contends, we see a number of reasons why we should not lie.
Firstly, lying has negative consequences for the person who lies. The mental accounting required to keep track of lies or misrepresentations of oneself is difficult. Liars, therefore, incur a mental cost that those who tell the truth avoid (p. 38). Lying also leads to shame, remorse and a loss of integrity (p. 40).
Lying also has negative consequences for our relationships (pp. 7-8). We are less likely to trust people who we know will tell white lies when it suits them. For example, if we discover that a friend will tell a white lie to get out of an engagement, we’re always left wondering whether they’re lying when they cancel an engagement with us. Such mistrust damages the relationship. A willingness to be honest — to avoid even white lies — lead to more gratifying relationships, Harris concludes, drawing on empirical evidence to support his claim (p. 37).
Certain ‘background truths’ (p. 19) or meta-ethical positions provide reasons why these individual and social consequences are negative. We want the best for ourselves and for our friends. Insofar as lying leads to a loss of personal integrity and undermines trust in our relationships, it leads to a sub-optimal view of ourselves and to sub-optimal friendships. Hence, it’s better not to lie.
Harris also discusses an important third category, what I’ll term negative ‘institutional’ consequences. When someone in a position of authority lies, their lies can undermine our trust in the institutions that they represent or to which they belong. We lose trust in political institutions when we discover that our politicians or bureaucrats have lied, Harris argues (p. 42). Citing the important case of now-debunked research linking vaccines to autism, he also argues that people can lose trust in science when they discover that individual scientists have lied (p. 43). The lies of one or a few politicians or scientists can undermine the public’s trust in an entire institution.
However, there are situations in which institutional representatives may lie without negative consequences — principally, Harris argues, in international relations. Sometimes national security requires a government to keep secrets from other states. Similarly, national security requires that we employ spies who must lie in order to obtain information from foreign governments and foreign nationals. But there’s an important difference between a government lying to another government and a government lying to its people. The background conditions or ‘truths’ that deem certain consequences of lying between friends to be negative don’t exist between states. We don’t want the best for other states, Harris argues (p. 44). Therefore, insofar as no negative consequences stem from lying between nations, he’s okay with it.
Harris provides a final reason why lying leads to negative consequences and is, therefore, to be avoided. Lying is epistemologically unsound. When you lie you are not relaying correct information about the world. Correct information is required to make informed decisions. By lying, you rob people of the opportunity to make informed decisions (p. 9). The same holds for false encouragement. False encouragement — for example, telling a friend that his or her work is good when it is not — Harris argues is ‘a kind of theft’ (p. 18). It robs the friend of time — time wasted on something they’re no good at; time that could be spent on other, more fruitful, endeavours.
So there you have it: Harris’s main reasons to avoid lying. As he sums it up, ‘Lying is, almost by definition, a refusal to cooperate with others. It condenses a lack of trust and trustworthiness into a single act’ (p. 47). ‘Every lie’ he adds, ‘is a direct assault upon the autonomy of those we lie to’ (p. 47). Lies are ‘the social equivalent of toxic waste’ (p. 48).
Having covered Harris’s main arguments, I’ll provide a quick objection before closing. It’s not an objection to Harris’s argument per se, more of an objection to his consequentialism.
As we’ve seen, Harris uses a consequentialist arguments to defend truth-telling, listing several reasons why lying leads to negative consequences. However, he then goes on to claim (p. 26),
A wasteland of embarrassment and social upheaval can be neatly avoided by following a single precept in life: Do not lie.
Here’s the problem with drawing categorical injunctions based on specific consequences. (It’s a problem with rule consequentialism.)
Through a series of everyday experiences and empirical research, Harris and others make a set of inferences about the link between lying and a sub-optimal sense of personal worth, lying and sub-optimal relationships, and lying and sub-optimal institutional standing. These inferences and a concern with the negative consequences of lying leads Harris to the ethical injunction, ‘Do not lie’. Now, conclusions of some sort must be drawn at some stage if an ethical system is to guide our action at all — if an ethical system is to give us some idea of how we should act. But rule-consequentialists like Harris need to recognise that how they formulate their ethical precepts alters the epistemic status of their ethics.
All conclusions drawn from empirical evidence are subject to the limits of induction. Imagine an ornothologist gathering empirical information about swans. After years of study, all of the swans that she has encountered have been white. By keeping up with all of the latest scientific journals she knows that all of the swans that her fellow ornothologists have encountered have been white. On this extensive evidence she concludes that all swans are white.
But we Aussies know that black swans exist. Our ornothologist’s conclusion was wrong. Importantly, however, she wasn’t wrong just because she and her fellow ornothologists didn’t have access to the relevant empirical information about the existence of black swans in Western Australia. She was wrong because the conclusion that she drew from the empirical information — that all swans are white — went beyond the limits of the valid conclusions that one can draw based on empirical information.
The conclusion that all swans are white is categorical. Categorical conclusions have to be deduced from axioms and definitions, or drawn from an exhaustive data set. But all sets of empirical information are limited, not matter how large the set. Therefore, a limited set of empirical information never warrants a categorical conclusion. To apply this reasoning to our example, no matter how much empirical information one gathers about swans, that information can never form the basis of a claim about all swans that exist. The information can only be used to make conclusions about all of the swans that have been observed. Now, if the set of observed swans and the set of all swans are the same then one could conceivably draw conclusions about all swans based on information about all the swans that exist. However, one can never know that the set of all swans that exists is the same as the set of all swans that have been observed because one never knows for sure that one has observed all swans. There could still be unobserved swans somewhere.
Harris makes the same mistake as our ornothologist when he derives the categorical injunction ‘Do not lie’ from a limited set of empirical observations about the negative consequences of lying. Categorical formulations of consequentialist ethical precepts go beyond the warrant of the empirical evidence on which they are based and are, therefore, invalid.
One solution is for Harris to qualify his precept, similar to how a scientists qualify the claims to knowledge that they make via induction. Our ornothologist salvages her conclusion when she changes it from ‘All swans are white’ to ‘All swans observed to date have been white’. Harris can likewise salvage his claim. His empirical evidence warrants the conclusion that ‘All instances of lying observed to date in which certain background conditions exists have caused more personal, social and institutional harm than good’. This conclusion, in turn, warrants the ethical precept ‘In situations similar to those observed, do not lie’. Though clunky, when presented in this form the ethical precept is recognised as only applying to cases that are similar to those from which the conclusions about the negative consequences of lying were drawn. As such, the precept stays within the limits of the evidence on which it is based. Therefore, the simple precept ‘Do not lie’ needs to be qualified if it is to retain the epistemic integrity that its empirical grounding implies.
The problem with rule consequentialism emerges at the precept-forming stage. At this moment, rule consequentialists lose an integral aspect of consequentialism: its concern for consequences as they occur in the real world. As soon as a precept is adopted, the collection of evidence ceases. The consequentialist precept ‘Do not lie’ is as categorical as the deontological injunction ‘Never lie’. And by acting on a precept alone, the person who was once a consequentialist risks becoming what I would call an inductive deontologist.
Deontology is an approach to ethics concerned with whether actions comply with ethical rules or principles, not whether our actions lead to good or bad consequences. A consequentialist would become an inductive deontologist if he or she was concerned with consequences insofar as they provided an empirical basis for ethical rules or precepts, but then closed off the evidence-gathering phase of their endeavour once they’d obtained that precept. This, I would argue, is exactly what Harris does when he adopts a categorical precept. And in so doing he becomes an inductive deontologist.
This closing off of the ethical system to further empirical evidence is a problem for the inductive deontologist because, unlike the deductive deontologist, the inductive deontologist has previously professed a concern with the empirical world. Once the evidence was in, however — once the conclusions were reached and the categorical precept obtained — the empiricism vanished. And so did any substantial difference between the ethics of the Harris’s ‘Do not lie’ and the deductive deontologist’s categorical injunction, ‘Never lie’. Both have a precept or principle that from this point on guides their actions. Both ethical precepts are closed off to further consequences — evidence that may very well disprove the precept.