Several theologians following Aquinas have said that when we say things like "God is good" that this must mean something different to when we call other things good; this is called analogical use of language.

For Aquinas, language can only be used analogically of God (from analogia, the Greek for proportion). Saying Mary is good and God is good shares some meaning, a proportion of the meaning, but not all the meaning. Aquinas uses a truly medieval example to explain. A good bull has a sleek coat, big muscles and a strong interest in cows; a good God would scarcely have these attributes! Nevertheless a good bull also produces good things (healthy urine and manure, high-quality semen and prize-winning calves) and does what good bulls are supposed to do, conforms to the ideal. In this we can see the proportion of meaning that could be shared between a good bull and a good God. God could also produce good things and fully fulfil his divine nature, not falling short in any respect. God being good in that he produces good things is known as analogy of attribution; God being God in that he perfectly fulfils His nature is known as analogy of proportion. (The Tablet)

To say that "good" means the same thing in regards to God and bulls is called univocal language, and is denied by many.

But this seems to me to conflate the sense of a word with its connotations. Obviously what it means for a hamburger to be good is different from what it means for a day to be a good day for flying kites. But I would say this is because of the contextual connotations, not an actual difference in sense of the word "good". According to the Natural Semantic Metalanguage Good is a semantic prime, one of the basic concepts of human language, and which cannot be usefully defined with other words or primes.

Is it fair to say that those who say that "God is good" is analogical are conflating sense with connotation? Did Aquinas and other medieval writers ever discuss the univocal/analogical divide with the Latin word connotō?

  • 1
    This kind of mixes Fregean and Millian terminology, sense and connotation are often used as synonyms. I think what you mean is literal vs metaphorical meaning, or semantics vs pragmatics, but that distinction is disputed by many modern authors, e.g. Davidson, so it is hard to call denying it "conflation". Scholastic authors did not subscribe to modern truth conditional semantics either, see Eco, Meaning and Denotation, meanings were "affections in the soul".
    – Conifold
    Aug 15, 2019 at 5:18
  • 4
    A detailed discussion of Aquinas vs Scotus on univocity is in Hall's book Thomas Aquinas & John Duns Scotus: Natural Theology in the High Middle Ages (freely available). It is not clear that either had such distinction in mind.
    – Conifold
    Aug 15, 2019 at 5:21
  • 1
    Definitely not "literal" vs metaphorical (I don't think it's a real division, everything is idiomatic, and many metaphorical meanings are part of the sense of words), but the division or lack thereof between semantics and pragmatics is something I had considered. I have read a little bit about Scotus and his position does seem better I think. So I'm more trying to make sense of the original univocal/analogical divide (which is still taught by Christians today). Aug 15, 2019 at 5:21
  • And thanks for the book suggestion, added it to my list. Aug 15, 2019 at 6:05
  • 3
    Barnes gives a survey on the uses of connotare in The Doctrine of Connotation and Denotation. It did not mean what you have it mean, neither for scholastics nor for Mill. For Ockham, for example, "white" denotes white things and connotes whiteness, i.e. it is close to what Frege called sense. They also used significare, apellare and denotare, none of them in your sense, see Eco. The use for "overtones of meaning" is colloquial and recent, univocal/analogical is the relevant terminology for that time.
    – Conifold
    Aug 17, 2019 at 5:16

1 Answer 1


The example cited to argue against "univocalism" -- "Good Bull" being very different from "Good God", is instead a clear case of equivocation. "Goodness" as exemplifying a "proper and healthy nature" is not morality, but comparative effectiveness in carrying out a role in this world. Whereas the efforts to avoid evaluating God on human moral terms, citing possible moral shortcomings in source document behavior, or aspects of our world, is s solely moral usage of "good".

Almost all words have multiple definitions, and equivocation is therefore a risk in many applications. This is particularly true in philosophy. The way to avoid an equivocation issue is to spell put an idea being considered, in more than one word, and therefore try to make the idea explicit and distinguishable enough to avoid equivocation fallacies.

For "Good God" in the "good bull" sense, it is possible that Zeus was a "Good God", as he oversaw the world, keeping it from getting too far out of harmony, and exercised his power -- patriarchally -- which he was an exemplar of. But his rapes, affairs, fickleness, and lack of concern for details of the world's operations -- would not make him a MORAL God, or even a very good manager. He instead exemplified the strengths and weaknesses of patriarchs, which was his role in the Greek Pantheon. This sense of Good is irrelevant to the issues that The Problem of Evil raise, and which an effort to separate God's Good form Human Good is a possible response to.

What Aquinas and subsequent theologians are trying to get at, is that a MORAL God, one that exemplifies Omnibenevolence, and the perfection of morality, might not behave identically to what we humans extrapolate from our own moral sense and experience. Note this is an argument that morality is subjective and situational -- which directly contradicts what much contemporary "conservative" moral theology claims.

As a logic claim, this point is valid. We have an intuitive sense of moral principles from our own experience with morality. We extrapolate from our moral experience to infer what an idealized morality would be. But there COULD be some situational aspect to objective morality, that we have not perceived because we are basically only in one (our) situation, such that the other Omni-characteristics change the nature of Omnibenevolence.

The issue for this logic claim, is that the whole reason to think about this question is to do empiricism. It is to evaluate the questions "Is this God as described, Omnibenevolent, and is this world compatible with an Omnibenevolent God?" And to answer this question, one has to assume a definition for omnibenevolent that is well enough understood to be testable. And the collection of "all possible exceptions to our inferred concept of Omnibenevolence, due to logically possible but as yet not understood situational conditions" is NOT testable. And the point of empiricism is to evaluate our actual models against the contingent world, and proceed to make judgments about the world, based on the ones that seem to match it. NOT to sit inactive and unjudging, because our models may not encompass all possible potential logic alternatives.

We CAN test our extrapolation of what Omnibenevolence appears to mean, based on our human experience of morality. And the problems this causes for thesim are the reason that most non-theists consider the Abrahamic God to be immoral, and think that the Problem of Evil refutes Omni-theism.

So YES, there may be some TBD unknown exceptions to omni-benevolence due to the other Omni aspects of a deity. But citing the possibility of other models or data that could overturn our current judgement applies to ALL empiricism, ALWAYS. And is NEVER justification to reject empiricism.

  • 1
    But good doesn't mean moral. It is but one connotation of goodness. Jan 20, 2022 at 22:33
  • 1
    For it to be equivocation I'd like to see specific evidence that the Latin word is polysemous, with separate senses of moral and proper/healthy applied and distinguished for all types of subjects, not just God. Jan 20, 2022 at 22:41
  • @curiousdannii -- that is an irrelevant point for equivocation. One cannot remove an equivocation by converting to a different language. YOU just articulated two different meanings, you understand the difference, making the conflation of two meanings an equivocation, whether one is using a poverty-stricken language like the fictional NewSpeak, or a more complex one.
    – Dcleve
    Jan 20, 2022 at 22:48
  • 1
    Sorry, but I don't think you're understanding my question. What I want to know is, from the perspective of modern linguistics, would it be accurate to say that those denying univocal understandings of God's goodness are conflating sense and connotation. Aquinas and others may have called it equivocation. Maybe equivocation means different connotations for the one sense. But at other times it surely means different (polysemous) senses. Would they say that it is univocal to call both a bull and a flower "good"? Or both a flower and a child? Jan 20, 2022 at 22:54
  • 1
    "YOU just articulated two different meanings, you understand the difference" No, I stated the difference between "moral" and "healthy" and "proper". Those words all have different senses. But they are also all connotations of "good". But those connotations don't mean that "good" has three senses. Jan 20, 2022 at 22:55

You must log in to answer this question.

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