In philosophy and science, we may wish to argue that some hypothesis 'explains' some facts we observe. In which case, we may lay out some assumptions for our hypothesis, in order to show how these assumptions entail the facts we wish to explain. (Kind of like the nomological-deductive model of explanation.)

I'm assessing 'God' as an explanation for the existence of the universe and considering various versions of a 'God hypothesis' (the assumptions required, deducing predictions from them, etc.).

What, technically, makes one assumption or postulate distinct from another?

What I want to be able to do is assess the 'complexity' of each hypothesis (by complexity, I mean the number of assumptions required to build it). Therefore, I need to reduce the assumptions down to their most basic form.

For example, here is a possible assumption in one of the God hypotheses:

  • There exists a disembodied mind.

It is possible to reduce the idea of a 'mind' into many different qualities, such as logical thought, the ability to have desires, the ability to have emotions, etc.

Would I have to reduce this assumption further? Why?

What is the structure of a "maximally basic" assumption?

  • 3
    There is a long tradition of judging "complexity" and "number" of assumptions in connection with the "law of parsimony", a.k.a. the Occam's razor, see e.g. Simplicity on SEP. Unfortunately, the "complexity" of the God hypothesis is itself the point of contention, theists see it as a single simple hypothesis that explains it all, non-theists see at as an extraodinarily complex hypothesis (without extraordinary evidence) that explains nothing, see Parsimony and God.
    – Conifold
    Commented Aug 13, 2018 at 23:52
  • Yeah, that was only an example within the context of my motivation for asking the question. The way I tend to formalise an understanding of parsimony is along the lines of 'how many' assumptions are made since the laws of probability dictate that we have a priori reasons to consider hypotheses that are more complex in this way to be less likely to be the case. Commented Aug 14, 2018 at 1:25
  • The question itself was actually about how to know when an assumption has been reduced to its most basic form. Commented Aug 14, 2018 at 1:26
  • 1
    'How many' is not very promising, one can always make it 1 by taking the conjunction, and assignment of "prior probabilities" is altogether a dubious endeavor. You should look at theories of explanation, parsimony is not the only factor that contributes to explanatory value, perhaps not even the main one.
    – Conifold
    Commented Aug 14, 2018 at 2:31

3 Answers 3


There is no such thing as a "most basic assumption". Lacan has expressed this in the notion that 'master signifiers are empty'.

(At this point, I did not intend to make a major digression. Please read whatever you want of this and then skip to the bolded 'Yet'...)

His motivating example seems to be Fatherhood. This is a notion that seems to be a basic biological fact, but is instead a social construct that did not exist in early societies. The roles of men as the protectors of specific sets of children is not uniformly associated with biological parentage -- such as in the Mabinogion and other very old epics (including hints in the stories of Abraham), where male privileges and training pass to the nephews of a mother's favorite brother, who is often an adopted friend. Our modern notion of fatherhood comes down from the more general notion of patriarchy, where men guard and guide collections of people, with the nuclear family being the smallest such collection. But not so many generations ago, in many societies the nuclear family's father was subject to his wife's father, or to his own mother, and did not control anyone, including his wife and children. We carried this notion of kingship downward over time, in a way that is not in any way simple. Etc. etc. There ends up being very little about fatherhood that is actually consistent and has meaning.

And in other ways as a 'basic idea' it is a knot of negotiations across time, which don't share anything -- our legal concept of fatherhood, at least in most US states, even proceeds in three directions -- biology, marriage and common law -- which do not agree upon what a father is. A child born to a legally wedded woman has her husband as its father, whoever's biological child it is, even if he abandons the family as a result. Marriage after the birth of the child may or may not confer fatherhood on the new husband, depending upon the level of involvement of the previous husband. A man who supports a child for a given period of time becomes its father even if there is no biological or marital link. Etc. etc. etc.

Yet 'the Father' seems simple and holistic as a symbol and is extremely compelling psychologically, even though it artificially binds up our species overall history of military rulership and male responsibility for acquiring resources from outside the tribe with the tendency of women to be more monogamous than men and to need close friends more intensely, and a bunch of other factors that just aren't simple.

Explanation is necessarily circular, with extremely complex ideas given the status of being simple by the force of culture or by appeal to shared emotional impulses, and extremely simple ideas becoming quite complex as one explores them, because they are only actually simple within a complex context.

This means that parsimony has to do not with absolute meaning, but with emotional resonance. Those concepts are simple which people can be expected to find compelling, and in general that means things that appeal to the broadest shared emotional level. One component of this broadly shared level is logic, so things that take elegant forms and are well-connected logically are often parsimonious. But beyond that, we have only Archetypes.

Concepts like power (safety) and abstraction (mind), and other things we consider to proceed from the religious notion of God (often The Father) are also very broad anthropologically -- but they proceed from a specific path to the notion of God, which succeeds in our shared social trends.

As writers like Starhawk have analyzed, you need to be wary of them, because 'sky gods', the kinds of gods that succeed in cities and other dense populations guarded by military and agricultural strength, are only dominant because those are the kinds of cultures that have taken control of the planet. In periods less given to oversimplification, alongside our God, the Western, Abrahamic traditions have often also had a feminine principle, from Baal down to Mary or the Shekinah. Cultures outside the West have, too. We find such threads less conceptually simple, but only because we have almost uniformly chosen a patriarchal style of life.

In many ways, the most parsimonious theory of God is Jung's: Everyone needs a parent, maybe two, and the ones that we understand too well, are not good enough.


How to know when an assumption has been reduced to its most basic form?

I will interpret this as how to know when an assumption is minimal and useful for making progress scientifically. When you have identified enough unique, orthogonal (independent) parts or properties of your hypothesized model that allow you to make specific, reliable (greater than random chance) predictions such that you can consistently recreate any effects that your model causes. I.e. you can design valid experiments or simulations that give repeatable results for a sufficiently large space of initial conditions, demonstrating that your model is robust and general, but at the same time gives specific predictions.

Sometimes that is difficult or impossible, so the next best thing is to know enough properties to make observations that confirms or denies your model, as in astronomy and evolutionary biology. These parts and properties must give rise to a unique process determined by your hypothesized model or cause. Science explains by describing properties and processes. The problem with minds is that they imply a hidden process that make choices. The effect (the observed universe) is what it is because of choices by god, according to some theologians. A choice is not a process or acceptable explanation unless you know all relevant internal mechanisms, history and factors determining that choice, as we know for human minds.

How does one judge “complexity” of assumptions for the purposes of “best explanation”?

As was pointed out in the comments, the Parsimony Principle. It's important to mention your goal and context of "best explanation" because there is no such thing as an objective "best", only best relative to a goal. The goal in science is to maximize descriptive accuracy and consistency with all previous laws and well-established models in science. As far as I can tell, in the context of theological arguments, "best" usually means consistent with bare human intuition and no accumulated history, knowledge or science, as a prehistoric human child might reason. It can also mean best for preservation of theological dogma or religious order, but never for objective descriptive accuracy or truth without extreme bias towards human special creation or purpose of the universe as a "gift" to humans.

An often overlooked aspect of parsimony is informational or descriptive complexity. If you use some word like "mind" you must describe and support with evidence all major parts both physical and non-physical. This includes internal mental "subroutines", functional units, properties and history in the same detail as you would for any other well-known phenomenon or model in science. An alleged creator of a universe must be at least as complex mentally as the only other known high intelligence, humans, for which we have exabytes of information that are critical in understanding behavior and decisions. So when weighing two explanations (mind vs. unknown natural processes or unknown combination of known simple natural laws) you must take into account the sum total of all information in a hypothetical library containing thousands of books that you would need to describe the hypothesized mind, just as we have for humans. To make it reliable science you must back up all of that information with scientific evidence derived from observations or experiments that put both cause and effect together at the same time in an unambiguous dependency.

There exists a disembodied 'mind'.

You can assume whatever you want for a philosophical argument but scientifically speaking, "disembodied mind" is an impossible self contradiction, given all we know about minds, how they work and what is necessary for a mind to exist. All known minds depend on matter or states of electromagnetic material fields. God is an extreme violation of parsimony because a mind is possibly the most complex thing in existence, short of social systems with multiple minds. Minds have complex sequences of states, modules, functions, behavior and interactions with environment.

As far as I can tell, the "simplicity" in theological arguments confuses informational complexity with linguistic or syntactic complexity. You can reduce the linguistic complexity of any arbitrarily complex concept by substituting phrases for single words until you eventually get to god-done-it. Science explains successfully because it is bottom-up, simple first. Top-down theological arguments fail because they start with an axiomatic "god-done-it" without defining all necessary words, actions or parts crucial for understanding. They fail to fill in the details of "how/when/where/why" of any decisions that this mind makes and they especially fail at observing the cause/effect together at the same time, as all other mechanisms in science.


From Occam's razor wiki :

Bertrand Russell offers a particular version of Occam's razor: "Whenever possible, substitute constructions out of known entities for inferences to unknown entities."

or, should you deem wikipedia "unreliable" (see comments): quote comes from : Logical Constructions. Metaphysics Research Lab, Stanford University. 2016.

  • my post IS a citation Commented Jan 23, 2019 at 7:22
  • Wikipedia is not considered a reliable source here. Also, an answer is supposed to include explanation and context instead of dropping quotes, as relevant as they may be.
    – Philip Klöcking
    Commented Jan 23, 2019 at 7:52
  • @PhilipKlöcking should you find my answer unclear, you are welcome to ask for further clarifications. I believe in conscision and dont believe in unnecessary verbiage. Commented Jan 23, 2019 at 9:06
  • We are looking for answers that do not require further clarifications in the comments, they should be added to the post itself. In this case, the post does not answer the title question in any apparent way.
    – Conifold
    Commented Jan 23, 2019 at 23:23
  • @Conifold no further clarifications were asked regarding the answer. I take it that none were required. My answer is a straight proposal for lowering the complexity in order to offer the best explanation, and therefore a straight answer to the question. Commented Jan 24, 2019 at 1:47

You must log in to answer this question.

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