According to George E. Moore, ethical claims all concern human conduct while philosophical ethics ultimately concerns itself with knowledge of what “good” is. Moore also believes philosophical ethics ought to concern itself with what is good instrumentally, or good as a means rather than good as an end, as a property. According to Moore, what is intrinsically good, or the property of “goodness” is not an analyzable property. For Moore, what “good” is, or “goodness”, as an individual property, is “unanalyzable”, or, undefinable. Therefore, any claim which gives a definition of “goodness” is attributing goodness to something, rather than identifying what goodness itself, as a property, is. Moore accuses those who make this error of committing the “naturalistic fallacy”. He believes that moral naturalists — philosophers who maintain that moral properties exist and can be objectively studied, through biology and sciences — are primarily responsible for this mistake. Moore thought philosophers committed the naturalistic fallacy when attempting to define “good” by moving from one claim that a thing is “good” to the claim that “good” is that thing. Moore thought one could not identify “good” with a thing one believes is “good”.
In order to test and determine whether an attempt at defining “good” is correct and not a concealed assignment is what Moore called the “open question argument.” Moore proposed that if “goodness” is a natural property, then there is some correct explanation of which natural property it is. For example, maybe “goodness” is the same property as “pleasantness”, or the same property as being “desirable”. Further, a correct property must be identified to fill in an identity statement of the form “goodness = __________”, or, “what is good is _________”. This kind of identity statement can be correct only if both terms on either side of the identity sign are synonyms for proficient speakers who understand both terms. Synonymy of the two terms is then tested through substitution of a term. Moore’s idea is that substitution of synonyms for one another preserves the original proposition that a sentence expresses. For example, using the sentence: “what is good is pleasant.” For this to pass Moore’s test, the sentence would have to express the same thing as “what is pleasant is pleasant.” Moore believed it was obvious that these two sentences do not express the same proposition. In thinking that what is good is pleasant, Moore thought one is not only thinking that what is pleasant is pleasant. According to Moore, there is an “open question” as to whether what is good is pleasant, and it can be understood when someone doubts the generated statement. However, there is no “open question” as to whether what is pleasant is pleasant, because this analytic truth cannot be doubted. Therefore, Moore thought that no substitution will pass the test. Therefore, there is no natural property of “goodness”. In other words, according to Moore and his open question argument, “goodness” is a non-natural property.
Objections to the open question argument include the fact that Moore assumes that an analysis essentially cannot be informative. It is also possible that “good” and some natural property Y have the same reference but different senses. If that were the case, “good” might refer to the same thing that “pleasure” does but still mean a different thing. Most importantly however, according to W. Frankena, Moore “begs the question”, as his premises assume the truth of the conclusion, rather than supporting it. Finally, Moore would run into the same trap as everyone else as he sought to prove that “goodness” is a non-natural property, but the open question argument can apply as much to that definition as it can to any naturalistic definition.