In his 2013 book Why Tolerate Religion?, Brian Leiter defines "religion" as follows:

  • Central beliefs that issue in categorical demands on action -- that is, demands that must be satisfied no matter what an individual's antecedent desires and no matter what incentives or disincentives the world offers up ("categoricity of religious demands").

  • Central beliefs that do not answer ultimately (or at the limit) to evidence and reasons, as these are understood in other domains concerned with knowledge of the world. Religious beliefs, in virtue of being based on "faith", are insulated from ordinary standards of evidence and rational justification, the ones we employ in both common sense and in science ('insulation from reason and evidence").

  • Beliefs that involve, explicitly or implicitly, a metaphysics of ultimate reality ("metaphysics of ultimate reality").

  • Beliefs that render intelligible and tolerable the basic existential facts about human life, such as suffering and death ("existential consolation").

Leiter explicated this definition in his book. Are there any weaknesses in Leiter's definition?

  • 5
    Correct me if I'm wrong, but you seem to be soliciting an open-ended discussion of how Leiter defines "religion". Our Q&A format isn't really conducive to such broad discourse - unless you can narrow down your question, it might be difficult to get what you're looking for.
    – commando
    Aug 10, 2016 at 4:41
  • Thank you. When I filled in my profile a few minutes ago, I referenced you. With regard to my question, I'm looking for a definition of religion. I thought I would start with Leiter's because I found it stimulating when I read his book. I wanted to see what others much smarter than I am had to say on the subject. Any suggestions on how to better scope the question? Aug 10, 2016 at 5:14
  • 3
    I think if you reword the question to Are there any weaknesses to Leiter's definition?, then it might be more objectively answerable. (but even then that seems pretty broad).
    – virmaior
    Aug 10, 2016 at 7:25
  • 4
    His definition excludes Christianity, because he doesn't seem to understand that true faith is a form of evidence capable of providing a foundation of certainty which exceeds the assumptions upon which the sciences are based. Apparently his definition only applies to humanistic religions, speculation and mythology.
    – user3017
    Aug 10, 2016 at 13:38
  • 1
    This definition cannot accommodate Shinto, which as I understand it does not do any of these things.
    – CriglCragl
    Mar 8, 2021 at 23:03

6 Answers 6


Any definition of "religion" that doesn't recognize it as primarily a sociocultural matrix is a bit suspect in my book. I would personally define religion as a set of defined values and ritual practices, attached to a community, arising from a combination of a developed theology and an accompanying set of folk beliefs, and typically having as an official purpose the development and advancement of a relationship with God and/or the divine (as noted in the comments, however, some "religions" are at least nominally non-theistic).

Mr. Leiter seems to be focusing entirely on the the beliefs portion of the religion, and without any distinction between the developed theology and the folk beliefs. Even here, however, I have some reservations. His first point --that religious duties override all other concerns --is true in theory, but often not in practice. It is a valid statement about religious commitments, but it does not necessarily match the reality of religion for the vast majority of adherents. His second point, that religious belief does not "answer" to evidence is formulated in two separate ways. The second, that faith is "insulated" from ordinary standards, seems to me like editorializing. There is a difference between beliefs that are outside of ordinary evidence and beliefs that are "insulated" from such evidence.

  • 1
    I could lose the part about 'advancement' and 'the Divine', and include Buddhism.
    – user9166
    Aug 10, 2016 at 18:26
  • @jobermark I thought about that, but the most common definitions of religion assume theism. However, I have edited to address. Aug 10, 2016 at 19:30

Reinforcing @ChrisSunami, this definition would admit delusions. There is no such thing as a personal religion, whatever New Agers want to think. The word itself is a reference to maintaining historical connection, the 're-' references the past, and '-ligio' is the word for tying (e.g. ligament or ligature.) A religion either binds you to the past, or binds your group back together when you come apart.

There can be a personal creed, but traditional analysis of religion often breaks it up into creed, cult, and guidance (magisterium). And of those the cult is by far the most important part. A religion is practiced, not espoused, as the first observation seems to say before it undercuts itself by focussing on categoricity over practice.

Religious cults are explicitly cultural and historical phenomena. Even when they are heretical departures from a culture's main tradition, they are attempts to adapt existing cultural requirements to new insights.

The definition tries to allude to cultishness in noting that the practice does not necessarily change in response to reality, and to guidance in noting that it 'solves' problems of life, but it needs to go outside of belief, and actually include them, in order to make sense in context when we are discussing actual religions.

At an extreme distance from this credal focus, as a Hicksite Quaker, there is officially nothing that I need to believe in order to fully take part in my religion, the same would be largely true of a modern Unitarian-Universalist. But there are things I need to do, at least occasionally, and there are people from whom I am supposed to draw inspiration and take guidance as individuals or as role models (whose lives point toward ways to solve the problems of life).

(Clearly, if I did not believe various basic things, the actions would cease to be meaningful or the guidance inspiring. But the beliefs that back them up can take a range of different forms.)


Religion is...

Religion is organised affirmation of faith

Faith is belief without evidence

Evidence is relevant claims that can be judged by another party to be true or false

The only weakness I find in Leiter's definition is that it is needlessly "chatty" and draws into it criteria that are not needed to define a religion, such as demanding that a religion "involve, explicitly or implicitly, a metaphysics of ultimate reality". I for instance find that religions like Buddhism risk falling outside the scope of that definition.


Categorical demands in action are a feature of any social contract, whether through law or religion. It is almost tautological to say that "a person within a group must conform to norms set by their group".. What is there to "the group", if it has nothing binding it? The rules of secular society fit this requirement, regardless of whether their origin is in a theological society.

Within experience, there is no thing for which ultimate evidence can be gained (can the limit be defined?). A hypothesis test with a 0.00% significance level is unattainable if the population and sample size are different (if they are the same, no new information is gained).. In this way, the scientific method does not answer to any Ultimate either - its validity cannot be 100% proven by itself after all. The secular world is still stuck as substantively similar to the religious world.

Moving away from whatever "the ultimate" may be defined as, within the conventional world religion is subject to analysis and change. It would be fair to say that a reasoned approach to evidence is hindered within religious thought due to an established and generally unchangeable metaphysical belief, but Judaism today is not Judaism during the first temple period. Islam today is not Islam before their philosophy touched Aristotle. Again, no different essentially from the non-theist world.

What world-view does not have a metaphysical position? The common sensical hides a metaphysical position, the irrational hides a metaphysical position. Any interpretation of Quantum Mechanics rests on some ontological and epistemological world view.

When you create any theory with a positive metaphysical position, in that you literally have to go "beyond" physics, you find yourself beyond a-posteriori thought, beyond the hypothesis test - beyond the scientific method. If you use reason to attempt to deduce a positive characteristic of Being a-priori, reason will negate it. But if you state that "metaphysics does not exist", reason guides you into contradiction again. The dialogue of Parmenides exemplifies this problem. The presence of a metaphysical position is not the presence of religion but the presence of human thought and language, which in its relative bluntness, creates something out of no thing without even a second thought. You could choose to drop metaphysics, as the scientific method does (the method does not "create" any ideas of itself, it only strips away unreasonable a-priori hypotheses)..

Nagarjuna did this for Buddhism too however, arguably "neti neti" (not this, not that) for the definition of Brahman within Hinduism does something similar, as does Tawheed in Islam and Tao within Taoism. The metaphysical is not connected to religion (or any thing at all really - it is beyond all that experience offers). In that it is beyond experience, it is the realm of faith - there is no religion within saying "All is One". Metaphysical positions in and of themselves cannot provide rational and complete rule-sets for the physical world. Religion (more broadly ideology) appears when it is said "All is One, and that One is X". X will not stand up to reason because the conventional world is perceived to be relational and not essential - from some frame of reference within experience, X will not be.. When ideology and experience stand opposed and the ideology is followed, that is religion. When they stand opposed and experience is followed, that is the scientific method. When there is no ideology?

Central beliefs that do not Ultimately answer to evidence and reason (or any aspect, however defined, of the Ultimate) cannot provide ultimate answers to the human condition. Any answer to the question of suffering can be sufficient from some frame of reference. To say that the theist fares better at this than the non-theist, is to assume that a theist has a better understanding of the metaphysical world, but "understanding" is meaningless here, as it is a conventional, physical notion. Again no substantial difference between the theist and non-theist position is visible.


Any set of laws, the resulting social and economic setting, as well as the emergent or directed norms generate a “religion”, a system of belief, a pattern of habits in individuals - a way of life. Any set of laws, social norms, any socio-politico-economic ecosystem is connected to, influences and shapes the individual by definition. Adhering to no religion at all still implies one lives by certain rules, that one acts in certain ways, adheres to certain norms and ideals, and lives confined within a socio-politico-economic system, defined by some rules. One then, through one's experience, begins to formulate certain beliefs, certain directions, certain goals. Thus one begins to develop and maintain within oneself, a religion, a set of beliefs about a directive power. One has rituals and customs and traditions and follows them "religiously", whether one's god is one's work or the state or money, or just one's own self.

Religion, thus, refers to what I would term as a “psycho-social fact” - a structure resulting from the natural function and interaction of the mind and social systems. It appears in groups, in any social setting. Religion is then neither normative or prescriptive, it refers to an observable phenomenon - it's inescapable and exists whenever a human being resides inside a group - a setting of which it is an inherent feature.

A similar concept is that of Governance. Governance refers to a category of social facts, namely the processes of interaction and decision-making among the actors involved in a collective problem that lead to the creation, reinforcement, or reproduction of social norms and institutions. Therefore, as a social fact, governance is neither normative nor prescriptive: it refers to an observable phenomenon. Nor is it limited to any time or space, as it is observable in any human society. Decision-making processes, social norms, and institutions are inherent to social life, allowing members of any society to live together and cooperate, even without a state.


Edit : Philosophy, "french definition" : <<< Set of conceptions bearing on the principles of beings and things, on the role of man in the universe, on God, on history and, in general, on all the major problems of metaphysics.>>> Question : Is "conception" dogmatic ?

Religion is "conception".

A RELIGION (or "ideology") is that "in what you believe" or that "in who you believe". As there are no clear definition of it (We can doubt everything ), all definitions can be plausible.

And it is not only "believe without evidence".

Example: "The love of money is a religion". What men do not do to always have more ... It is not for nothing that we have said: "The love of money is the root of all evil". Not money itself. False ???

So there are as many religions as you want.

  • 1
    This answer is rather dogmatically worded. Unless you can revise it in the form of an argument, it will probably be deleted.
    – Geoffrey Thomas
    Sep 11, 2021 at 16:11
  • Ok. Religion IS dogmatic. Then no need to ask "what it is" !
    – Antonio51
    Sep 11, 2021 at 20:38

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .