[1] Quine, W.V. & Ullian, J.S. (1978) The Web Of Belief. (2nd ed) New York. Random House. p 11.

[2] Even if that argument is not sound, and there are other ways to correctly justify a moral theory, this book can be viewed as an investigation of the epistemic justifiability of action-guiding theories.

[3] Works by such writers are mentioned in the References/Bibliography at the end of this book.

[4] In philosophy, ‘prescription’ means a statement telling you what you (allegedly) should do. Many philosophers use ‘normative’ instead of ‘prescriptive’. I prefer ‘prescriptive’ because ‘normative theories’ suggests that the values the theories advocate are ‘norms’, i.e., normal — and this is often not the case.

[5] A moral skeptic doubts whether any moral theory is true or close to truth. A moral nihilist is more than just doubtful. A nihilist is sure all moral theories are false.

[6] Various problems with the notions ‘sufficient’ and ‘insufficient’ are discussed below. It can be far from clear how much evidence is sufficient.

[7] This is because the book’s primary aim is to develop an epistemically justifiable theory applicable to all intended practices. So the theory must be sufficiently practicable for this. It must give us and be based on practicable knowledge.

[8] We can use Newton’s physics to land on the moon. From the viewpoint of the truer or more generally true Einsteinian physics, Newton’s physics are epistemically questionable. If the scale is the whole universe, or galaxies, Einstein’s physics is much more accurate. But on the comparatively small scale of the Earth-to-moon distance, Newton is about as accurate as Einstein. So, for purposes such as moon-landings, Newton’s physics is epistemically sufficiently practicable. And it’s easier.

                There are controversies concerning such issues. E.g., see Standford, Kyle B. “Pyrrhic Victories for Scientific Realism”, J. Philosophy, vol C. 11. Nov 2003. pp 553-572, and that article’s references. However, various arguments below seem to plausibly resolve the controversies, at least regarding the book’s focus on practicable knowledge.

[9] This distinction relates to the ‘is-ought’ issue, discussed soon.

[10] Moral theories claiming this are criticised soon. One criticism here:- It is plausible that longest-term happiness could be maximised by having, for all future generations, e.g., only one racial type, the presently most numerous, say Chinese. This would avoid the great unhappiness involved in racial conflict. If true, is this evidence that all not of that race ought to be killed (painlessly) or not reproduce? Similarly regarding having only one religion etc.

[11] E.g., one theory claims that pleasure or happiness is what we ought to always only aim to achieve; other theories deny this, claiming instead, e.g., that being responsible, e.g., regarding child-raising and contributing to society, which often involve unpleasurable sacrifices, is most important.

[12] The discussion here is perhaps an over-simplification regarding knowledge in general, though not regarding morality compared with science. Analytic statements include ‘A black cat is black’ and ‘1+2=3’. They are reducible to ‘X is X’, which is circular. E.g., ‘A black cat is black’ is expressible as ‘A black thing is a black thing’, i.e., X is X; and ‘1+2=3’ is ‘1+1+1 is 1+1+1’, i.e., X is X. Some say we can have analytic knowledge, but there are some controversial issues here. The present point is merely and plausibly that we do not need to look outside of an analytic statement for evidence that it is true, i.e., ‘X is X’ just is true, formally — but this is no help in justifying any one morality among the many. ‘X is X’ is abstract, and formal. Morality concerns non-abstract content in that it refers to concrete things such as observable actions, one’s self-observable conscious intentions, and so on. Abstract, analytic statements are no help in investigating whether a moral theory’s content is (close to) the truth. This is because (1) ‘A green cat is green’ is in an analytic sense true, but this is separate from whether it is true there are green cats, in observable reality; and an equivalent separation applies to moral arguments; and (2) if an analytic statement, as in ‘X is X’, could justify non-abstract content, it could justify contradictory things. (This would be like the circular ‘X is right because X is right’, where X could be ‘murder’ or ‘not-murdering’, and so on.) The issue of using an abstract, formal statement in morality is discussed further later, when Kant is discussed.

[13] Dark reddish-purple is the next most useful colour, as in some Prunus trees.

[14] Part II argues that a type of fairness (and much else) can be indirectly epistemically justified. As a moral concept though, fairness is unjustifiable.

[15] The standard involves much more than simple observation. This is discussed further later.

[16] Some philosophers define ‘good’ and ‘right’ as meaning the same thing. Others define them as meaning different things. For the purposes of the present discussion, any of such definitions can apply.

[17] This raises the problem of induction, which is discussed later.

[18] We simply must make choices regarding such things as whether to kill ourselves or others, to act helpfully or unhelpfully, be selfish or unselfish, to care for children or not, or for this child more than that child, to lie or not, to eat meat or not, to smoke or not — and so on. Each such choice is a issue in the how-should-one-live sphere. Each such choice means implementing an associated theory, e.g., a theory advocating that we do not kill ourselves, or that we be helpful in certain situations. If, e.g., we at least implicitly prefer (choose) to be alive rather than not alive, we have an at least implicit theory here — the thought that life is more justifiable than death.

[19] Here I might seem to be holding a moral belief, namely that conflict, war, suffering and so on ought to be avoided. This possibly apparent inconsistency with the book’s anti-moral position is only apparent. Part II argues that there are other, non-moral, epistemically justifiable reasons for avoiding war and so on.

[20] There are complex problems here. A simplified answer is:- Young children are exempt from criticism here. If adults are aware or informed that and why their moral beliefs are questionable, and that they individually have an epistemically justifiable responsibility to investigate the epistemic justifiability of those beliefs and thereon-based acts, then they are fully responsible here. Education is crucial here. (Even young children can begin to understand such issues.)

[21] As Part II will make clear, here I am not beginning an argument allegedly justifying a rigid, authoritarian society. Quite the contrary. E.g., I’m seeking authority for total individual epistemic autonomy (freedom).

[22] I’m using ‘god’ as a non-sexist term, to mean a supernatural being which could have any or both or no gender(s).

[23] By ‘X is plausible’ I mean there is a significant possibility that X is correct. Here there seems to be insufficient evidence to epistemically justify X, but there is more apparently supporting evidence for X than against X, or about the same possibly-significant degree of apparently supporting evidence for X and for some alternative(s). E.g:- Imagine a new disease, the cause of which we have no idea. Going on all the evidence concerning all other known causes of diseases,  initially equally plausible hypotheses could include: bacteria, a virus and pollution. Implausible alleged causes include: witchcraft, demons and immorality.

[24] E.g., consider different views concerning the historical fact that Europeans were victorious over the indigenous populations of America. Agreeing on exactly the same fact or true ‘is’ here, some persons believe that the invasion and conquest ought to have occurred, but other persons believe the opposite. A person who has the moral belief that the European conquest of America ought to have occurred, does so ultimately because the person has a moral belief stating or implying that the European conquest of America ought to have occurred. (E.g., the foundation-belief that Europeans tended to be Christian and that Christians ought to rule non-Christians, or that the (allegedly) white (actually pinkish) races are superior and ought to rule the world, implies  that the  European conquest of America ought to have occurred.) A person who has the moral belief that the European conquest of America ought not to have occurred, does so ultimately because the person has a moral belief stating or implying that the European conquest of America ought not to have occurred.  Here there is obviously problematic circularity. This involves the same fact being interpreted via a moral belief. The fact does not imply any moral belief: there is no logical or epistemic connection between them. It is the person’s moral belief which causes the person’s moral interpretation of the fact. The fact does not cause a moral belief, or mean that the moral belief is also a fact, or true.

[25] Hume stressed that this gap exists, and then seems to argue that his theory fills it as far as possible.

[26] Some creatures, e.g., some fish and snakes, do normally kill and eat as many of their young as they can catch. Presumably there is a gene motivating this, or no gene motivating care of their young. Such young do tend to pass on their parents’ genes — simply because the genes give parents the ability to produce newborn which are so numerous, fast-moving and independent that most escape their parent’s jaws. The normal human desire to care for children was presumably naturally selected because newborn humans are very unlike those newborn.

[27] Part II suggests that something like this common core can be indirectly epistemically justifiable. It cannot and need not be justified by a moral theory.

[28] Divergible prescriptions are those a society can have which are very different from another society’s in the same area, and, all else  equal, that divergence will not affect the survivability of the society. E.g., in the area ‘dress codes’, suppose one society teaches that bare shoulders are immoral but bare knees are permissible, and another society teaches the opposite; and this divergence has no affect on the survivability of either society. The difference then, would not concern these societies’ common, survival-necessary morality. The distinction between a common core and divergible morals is only rough: it may be very difficult to correctly categorise some moral rules. But that does not matter for present purposes.

[29] The  event ‘moral beliefs exist’ is a fact, an ‘is’. A true explanation of that event is an ‘is’. The belief that we ought to have and act on moral beliefs is an ‘ought’. I’m arguing that there is an epistemic gap between that explanatory ‘is’ and that allegedly justified ‘ought’.

[30] An alternative explanation is that there is (are) some god(s) who egotistically put a god-worshipping part in our brain — so that we can worship the god(s). However, this and similar explanations lack objectivity: there is no evidence to epistemically justify any such explanation for us. And, if a specific god created that brain part, why create it so as to allow humans to worship innumerable other very different gods, and human charismatic leaders? There is ample evidence that natural selection occurs, and it seems highly plausible that evolutionary theory can explain brain events which bind a gene-protecting group together via beliefs they gullibly feel confidently comforted by.

[31] Instead of ‘Society of Ends’, many translators of Kant use ‘Kingdom of Ends’. I’ll use the former. Kant used the word ‘Reich’, which can mean many types of State or society. If he meant ‘Kingdom’, I think he would have always used ‘Koenigreich’, which specifically means ‘kingdom’. He did not. With this, Kant argued for a democratic society, with every individual having the maximum freedom consistent with every other individual having that freedom — not a king ruling over less free subjects.

[32] Hobbes, T. Leviathan. Ch XIII

[33] Kant’s theory often directly concerns ‘maxims’ rather than ends, but, as will be explained further below, each maxim implies an end. E.g., the maxims ‘Never lie’ and ‘Help others’ imply, respectively, achieving the ends ‘Lies are never told’ and ‘Others are helped’ or, at least, ‘I intended to help others’.

[34]  Regarding the problem of induction which arises here,  John Worrall makes amusing insightful comments. See Worrall, J (1989) “Why Popper and Watkins Fail to solve the problem of induction.”, Boston Studies in the Philosophy of Science, 117: 257-96. The problem is discussed further below, later.

[35] Calling this ‘war’ instead of ‘murder’ is a verbal delusion used to help make this moral belief acceptable.

[36] This change could occur via evolution, perhaps related to the fact that people now normally have an urban,  technological, mass-society environment, not the non-urban, technologically-primitive, small-tribe environment we evolved in. Or, Earth could be irradiated by something changing most people’s emotional neurophysiology or the common-morality gene(s) passed on to the next generation. So any universally accepted and universally applicable intersubjective theory, say Hume’s, could be replaced by a different intersubjectivity. This replaceability and contradictability would not apply with an objective or epistemically justifiable theory, as in science. (This is expanded on in Part II.)

[37] George Couvalis, Flinders University, South Australia, mentioned this point to me. Again, most ideas in Part I are not mine.

[38] Arguments such as this have been round since the ancient Greeks. Aristotle argued similarly, and I’ve used his ideas here as interpreted by John Hospers. (Hospers, J (1973) An Introduction To Philosophical Analysis. Routledge & Kegan Paul.)  Again, most points in Part I were made by other writers.

[39] Two positive (or negative) poles of a magnet do not attract each other. But here gravity still applies, though a magnet’s force can merely be stronger. Such issues are discussed further later.

[40] This example, and the just-mentioned one concerning murder, are mine (though Kant may use them in a minor text I’ve not read) . However, his theory implies them. The next examples, concerning selfishness generally and lying, are Kant’s.

[41] By ‘will’ here, Kant means something like ‘coherently or correctly imagine you’ve commanded’: you are the legislator of X in that you would cause X to be universally done if you could. This relates to the above point that, for Kant, a justifiable act is one which we can know (correctly imagine) is universally applicable, though it might not ever be universally applied (done). As suggested above, and will be explained further shortly, ‘will’ has nothing to do with being ‘willing’, i.e., being happy to do something. ‘Will’, in Kant’s most general categorical imperative, involves being able, without contradiction, to imagine something as universally applying, i.e., as a law. Kant rules out emotion, i.e., willingness and unwillingness, from laws.

[42] Problems regarding this definition are discussed soon.

[43] E.g., in Metaphysics of Morals AK (RPA) 6:224. (Gregor (1996) transl pp 16-17) (Gregor uses ‘AK’ to mean the Academy’s editions of Kant’s work, i.e.,  the Royal Prussian Academy’s editions, whereas I use ‘RPA’ for the same.)

[44] Because universal applicability and non-contradiction are inextricable regarding epistemically justifiable principles, as in scientific laws, the situation regarding morality or separate maxims shows that what  Kant calls ‘moral laws’ are not epistemically justifiable, and not laws.

[45] A farmer told me he saw this happen to one of his flocks of sheep.

[46] Groundwork RPApp 421-424 (‘RPAp’ means ‘Royal Prussian Academy page’)

[47] Rachels, James. (1999) The Elements Of Moral Philosophy. 3rd edition. McGraw-Hill. pp128-9.

[48] Some try to save Kant here by arguing that Kant’s basic principles are correct, but that he applies them to examples he should have avoided —  because examples such as that involving a would-be murderer are too problematic, and are argued to not fit with his correct, basic principles. (George Couvalis, Flinders University, South Australia, informally outlined to me this attempt to save Kant.) However, this attempt does not work. E.g:- A basic principle here, Kant says, is that there are only two types of law: negative (prohibitions, i.e., ‘Don’t do X’) and positive (commands, i.e., ‘Do Y’). An associated principle, Kant claims, is that a negative law applies regarding lying because a negative law is defined by its opposite being inconceivable. As argued in the previous chapter, ‘Always lie’ is inconceivable. So, according to Kant, the opposite of ‘Always lie’ is ‘Never lie’, and this is hence a negative law. Yet that associated principle is simply asserted. It seems ad hoc, lacking evidence. We cannot know it is correct or a fundamental truth. Why, e.g., are there not only positive laws, as in ‘Always be truthful’? Negatives can be stated as positives, and ‘Never be truthful’ is inconceivable. Even if that was not an epistemic problem, it is still the case that ‘Always lie’ is inconceivable, so Kant has applied his principle here correctly. But, plausibly, an opposite of ‘Always lie’ is ‘Don’t always lie’ — and ‘Don’t always lie’ can mean the partly negative, partly positive, ‘Don’t always lie, but do lie to achieve X’ or ‘Always be truthful, except to achieve X’. And  X can be many different, conflicting universalisable things, including ‘save innocents from murderers’ and ‘help murderers’. Kant’s UCI gives us no way to know what to do here. Other issues here are discussed later.

[49] E.g., in Mill, J.S. (1843) A System of Logic

[50] Kant tries to argue that ‘Don’t help others’ could not be a specific categorical imperative because no rational being would wish to not be helped when the being needed help (Groundwork, RPAp 423.); therefore a rational being would help others, which is universalisable. Kant says that ‘Don’t help others’ is universalisable, but that to obey this would contradict the (alleged) help-needing nature of a rational being. Hence, says Kant, this qualifies ‘Help others’ as a positive law, and, hence, an imperfect duty.  However, here Kant sneaks in  moral concepts he often insists must be ruled out. He sneaks in the notion that a human would wish to be helped when needed, and the notion that this (alleged) ‘is’ implies that the (alleged) wish ‘ought’ to be acted on. So here he has an is-ought problem. And elsewhere he rules out any specifics of human nature as a basis for morality. This includes human dependence on others, e.g., for food. A ‘rational being’ is not necessarily human, he points out. A rational being is simply something capable of reason, as in a computer, or a conceivable rational god or alien — each of which does or could lack human needs, dependence and wishes. E.g., an alien, once born, could conceivably be independent of others by photosynthesising its own food, and so on. Kant elsewhere also rules out what persons wish,  or would be willing to do, as a basis for morality. As argued above, Kant points out that if something is objectively right, it is irrelevant if we wish for or like it or not. The UCI only implies what is objectively right, insists Kant. Wishing to be helped sometimes is irrelevant here. Besides, even if a rational being as such, namely as in a computer, could wish, why would it wish to exist or survive? These issues are discussed further, later.

[51] Shortly, in another footnote, a problem regarding this prioritising is discussed.

[52] A specific definition of ‘innocent’ would be needed. I’ll ignore the problems here.

[53] This prioritising, this reversal of what Kant advocated, would also occur if one interpretation of  the criterion for what Kant calls ‘positive laws’ was given priority over the  criterion for what Kant calls ‘negative laws’. Here, consider a rational being which also has a (human?) nature which unavoidably wishes to help protect innocents, along with and via self-consistently wishing itself to be so protected. This being could not will the universalisation of a maxim such as ‘Never lie’ here, because doing so would contradict its nature (that wish). So, here, ‘Never lie’ would be overruled. As argued above (and below), there seems no epistemic justification for Kant giving one criterion or alleged type of law priority over another. Yet, if so, there would still be conflict in practice, over which to give priority, and whether the priority is absolute, or relative to different situations. Kant could still give no epistemic justification for one choice here rather than the other.

[54]  E.g., Kant says “[With a wide, i.e., imperfect, duty] the law ... leaves a playroom for free choice ..., the law cannot specify precisely in what way one is to act and how much one is to do .... But a wide duty is not to be taken as permission to make exceptions ... but only as permission to limit one ... duty by another (e.g., love of one’s neighbour in general by love of one’s parents).” Metaphysics of Morals AK (RPA) 6:390. Gregor (ed & transl) p 152 . Also:  “[One] ought to sacrifice a part of [one’s] welfare to others ... [but how] far [this sacrifice] should extend depends, in large part, on what each person’s true needs are in view of his sensibilities, and it must be left to each to decide this for himself ... [regarding] this duty[,which] ... is only a wide one.” (Here Kant recognises the area/maxim problem, but gives no clear solution.) In sum, Kant says we are not to perform an imperfect duty in every relevant situation, i.e., universally. E.g., even though it is universally the case that one could help someone else, Kant says one need not help others in every situation where they need help, because one is sometimes also obliged to fulfil one’s imperfect duty to do things just for oneself. And vice versa. We often cannot perform two such duties universally: they would conflict. Further here, it is highly problematic to know what “each person’s true needs are”. This is only resolvable via having a single, epistemically justifiable end, because  there are certain things which are true needs in relation to achieving such an end: everything which is justifiable except that end would be a means to that end. Otherwise, persons can conflict regarding what they believe are true (justifiable) needs. E.g., each of many conflicting religions tends to see belief in and support for itself, and disbelief in and opposition to other religions, as true needs. A ‘need’ is a need  for something, X, and a need for X is relative to the end, X, or to some further end because X helps achieve that end.  X is only a need from a perspective involving an end. Kant gives no epistemically justifiable end here (or anywhere). E.g., humans need food in order to survive, but Kant gives no coherent, epistemic justification for the survival of all, or any, humans.

[55] This point assumes that Kant says laws imply duties, i.e., that imperfect duties are not in an area separate from lawful acts. (Such separation would involve the just-mentioned area-problems.) However, the UCI says ‘Act only on a maxim you can will to be a universal law’, and Kant restates this as “I am never to act otherwise than so that I could also will that my maxim should become a universal law”. (My emphases.) (Groundwork, RPAp 402). So, there is no intended act not covered by a law. And Kant says we ought to do dutiful acts.  All justified acts then, are lawful, says Kant. Duties, which, as Kant says, are inherently justifiable acts, must be implied by a law, for Kant. Otherwise, Kant would not see duties as justifiable: they’d not be duties. Kant says duties are implied by laws. E.g., ‘Never lie’ implies a duty to never lie.

[56] Kant says the basis for distinguishing perfect from imperfect duties derives from the (alleged) fact  that negative laws imply perfect duties, and positive laws imply imperfect duties. Allegedly:- Perfect duties are those where we cannot conceive of a universalisable maxim which contradicts the duty. E.g., we cannot conceive of or correctly imagine universal lying. Imperfect duties, he says, are those where a maxim contradicting the duty can be conceived of as being universally done — but the nature of rational beings would be contradicted if the duty was not fulfilled. E.g., humans could universally obey ‘Be unhelpful to others’, but, Kant says, it is the nature of a rational being to sometimes need help — and any rational being would obey ‘Help others’ because the being would wish to be helped when the being needed it; and so on.
             I’ve discussed the problems here above, arguing that Kant’s basis for distinguishing between perfect and imperfect duties is too problematic. Besides, the present discussion relates them to universal applicability, which Kant stresses is the defining nature of all laws. Hence all duties, which Kant says are directly implied by laws, (as in the (alleged) law, ‘Never lie’, implying the duty, ‘Never lie’,) should be universally applied (obeyed). The discussion also fits with the UCI, which he says is the most general categorical imperative. This means the UCI’s criterion of universal applicability would apply to both perfect and imperfect duties. So both should be universally applicable, period.

                Further, Kant’s distinction between perfect duties and imperfect duties involves him giving priority to perfect duties. So, allegedly, in a potential conflict between a perfect and an imperfect duty, the perfect should be done. But this prioritising seems ad hoc, a mere assertion, without evidence. With this, any alleged negative law can be expressed positively, as in ‘Never lie’ being expressed as ‘Always be truthful’. This makes it a positive law, like ‘Be helpful’ or ‘Protect innocent lives’. So any priority given to a negative law and implied duty over a positive law and implied duty, i.e., the perfect over the imperfect, due to the alleged negativity, disappears.
             Also, if perfect and imperfect duties are implied by laws, namely morality’s equivalent to laws of nature, then to claim that one type of law and hence duty, e.g., an alleged negative type, has priority over another, is to contradict how laws of nature work. In nature, one law, say the law of friction, does not step aside and allow some other law(s) to have priority and be the only one(s) applicable. If nature’s laws worked thus, then, e.g., the law of friction,  which describes how brakes stop a car, would sometimes not apply. Then, via the laws of motion, cars would propel us unstoppably forward. In sum, according to a fundamental notion in Kant’s theory, a claim that one type of law has priority over and should/can replace another is merely a claim — without evidence. All the evidence there is suggests otherwise.

[57] Kant says we have a duty to perfect “any capacities whatever”. Gregor (transl & ed)  Metaphysics of Morals. p 154-5. AK 6:391-3. To get them all perfect seems a full-time (universally applicable) task.

[58] I’ll argue that, to the degree that human knowledge is incomplete, to that degree we’ll have unavoidable problems deciding a justified balance among actions in any (sub-)area. E.g., suppose ‘Eat only a balanced, healthy diet’ is justified. Present dietary knowledge and related health knowledge cannot tell each person the precise balance among proteins, carbohydrates, water, vitamins, fibre etc which they should eat at each meal. However, here we can have approximate knowledge, or get close enough to the truth, so that we have sufficiently practicable knowledge to normally keep people reasonably healthy. In many areas we can have such sufficiently practicable knowledge — and learn more.

[59] Groundwork RPAp 436. (‘RPAp’ means ‘Royal Prussian Academy page’)

[60] Kant’s rational beings as such could inherently only be motivated to achieve an end which is or is equivalent to a rational principle. Kant’s rational principle here is the UCI. This is insufficiently motivating for humans, so something equivalent but more motivating is needed, he suggests.

[61] Groundwork RPAp 427-430. Note that this law is expressed in a way which is both positive and negative. It can also be expressed as ‘Never treat persons merely as a means; always treat them (also) as an end.’. This positive-negative mix and reversibility is another example of  certain  problems discussed above (and in the Appendix), e.g., regarding Kant insisting there are only two types of laws, one positive, one negative.

[62] ‘Always treat persons as ends, never merely as a means’ would be just part of what the UCI implies, namely an imperfect duty, according to one interpretation of Kant’s criterion for a positive law and hence imperfect duty. If a rational being also has a nature which involves it not wishing to be treated merely as a means, this being could not will the opposite of that maxim because doing so would contradict its nature. Yet, if it is an imperfect duty, the ‘always’ and ‘never’ have to go, says Kant, because imperfect duties are flexible, not applying universally. So this would  be further problematic. And, if that imperfect duty is equivalent to the UCI, this would mean that either the UCI would have to stop applying universally, or this imperfect duty would have to apply universally and hence become a perfect duty.

[63] Groundwork RPAp 436.

[64] Groundwork RPAp 460-61

[65] Kant claims that humans have a unique emotion, a respect for rationality. See, e.g., Groundwork RPAp 400. Later, when dealing with a different issue, I’ll imply that this claim is correct, with qualifications.

[66] Groundwork, RPAp 428-9

[67] E.g., to always act on ‘Never lie’ is to achieve perfect truthfulness.

[68] Such characteristics tend to be aspects of common morality. So the present point can be seen as using a moral argument against Kant’s moral argument here. This implies that Kant’s argument is epistemically questionable in that an opposing argument, which is also from a moral standpoint, is more or less as plausible or appealing as Kant’s. This would be problematic. However, later I’ll argue that such characteristics can be advocated by an epistemically justifiable practical theory.

[69] E.g., Groundwork, RPAp 393-4 and 397-401.

[70] If ‘person’ is defined as a rational being as such, or only the rational aspect of a human, recent points show this would also be too problematic.

[71] In that we can know anything, including that oneself exists, we are rational and hence at least implicitly necessarily conceive of ourselves as rational. Yet this does not mean we also necessarily conceive of ourselves as ends-in-ourselves.

[72] Kant sees such statements as ‘merely empirical’, i.e., based on the observable. He contrasts the empirically objective with what he sees as the objective according to pure practical reason, which ignores the empirical. (‘1+2=3’ and the formal principle of non-contradiction, ‘Not both X and not-X’, are types of pure reasoning. So is the UCI.) Kant says that pure reason is what’s needed to discover the true or objective morality. He says empirical objectivity merely shows us what we do, but pure reason shows us what we should do (i.e., via the UCI). However, Kant’s claim that all rational beings necessarily conceive of themselves as ends-in-themselves is the claim that they do necessarily conceive of themselves as ends-in-themselves. So it is equivalent to an empirical claim. Besides, Kant suggests that self-observation is needed to show us what we do as purely rational beings; and self-observation is plausibly a type of empirical objectivity, because observation is the basis of empirical objectivity. Further, much of the above concerning Kant seems to show that pure practical reason, involving only the formal, as in the UCI, cannot unambiguously or epistemically justifiably show us what we should do. In sum, Kant’s concepts here seem confused.

[73] ‘Merely’ tends to have a negative connotation. ‘Only’ can avoid this.

[74] Groundwork, RPAp 423

[75] Kant defines an imperfect duty as the type of duty implied by a positive law, and he defines a law as a rule valid for all rational beings, not just for human beings. This means that notions such as ‘A human would wish to be helped, when needed’ are not part of Kant’s definition of an imperfect duty. Hence those concepts are sneaked in. Similarly:- He says that a positive law and hence imperfect duty are defined by the maxim opposite to the law being conceivable and universalisable, but that universalising the opposite maxim would contradict the nature of a rational being. Humans, with their dependence on help, do not fit the definition of rational beings as such. Besides, again, Kant’s definition of a positive law, with its implied alleged imperfect duty, as opposed to negative laws and implied perfect duties, seems ad hoc. Why shouldn’t all laws be of the same type? Kant stresses that moral laws are morality’s equivalent to nature’s laws, but there do not seem to be both positive and negative laws of nature, defined in a way equivalent to how Kant defines (alleged) positive and negative moral laws and associated duties.

[76] E.g., Groundwork, RPAp 425

[77] This was discussed in a previous footnote in this chapter. Humans who are unconscious do not have a rational aspect while unconscious.

[78] The only type of moral theory escaping this problem is one which states all persons ought to die, e.g., leaving non-human life-forms to run the planet better.

[79] Positively, freedom means freedom to do something. Negatively, freedom means freedom from something. In practice, the negative and positive go together, like two sides of a coin. E.g., suppose, when Fred turns 18, he becomes free from the legal restrictions on drinking alcohol when under 18. This means Fred is now free to legally drink alcohol.

[80] Groundwork RPAp 446-7

[81] Such points are made by Thomas Nagel, Cohen and others. See, e.g., Cohen, G.A. (1996) “Reason, Humanity And The Moral Law” in Korsgaard, C (1996) The Sources Of Normativity. Cambridge University Press  (Paperback edition.) 167-188 and Nagel, T (1996) “Universality And The Reflective Self” in the same book, pp 200-209

[82] Groundwork RPAp 447

[83] In a discussion in the Kant Reading Group, Adelaide University,  2003.

[84] See, e.g., work by Christine Korsgaard and Onora O’Neill. Works by them and other modern Kantians or neo-Kantians are in the References.

[85] A specific scientific practice aims at discovering truths about something, e.g., about fruit-growing or bomb-making. The specific end of a commercial or political director of the research here is, say, truths regarding how to grow more fruit or kill more people. Yet in that the science as such here is inherently aiming to be epistemically justifiable as such, science’s end or aim is knowledge, as opposed to the ends of different commercial, political etc directors of scientific research. Truth or knowledge as such is the universal, general end at which all scientific enterprises inherently aim. If we abstract from all the specific contents of each confirmed scientific theory, namely truths regarding fruits, bombs, planets, diseases and so on, we are left with the notion, ‘truth as such’. (Note that this is rather different from Kant’s claim, discussed in Part I, that if we abstract from the contents of all scientific laws, we are only left with the form ‘universal applicability, and, with this, non-contradiction’. I’ll argue that the former abstracting includes but goes significantly beyond Kant’s, taking us to an epistemically justifiable practical theory.)     

[86] Here lies one of the recently-mentioned problems regarding the specifics of this very abstract discussion: specifically which truths should one aim at or practise? Again, this problem is dealt with soon, when the context for doing so has been better developed.

[87] This is obvious to us, but not to all through history. Bertrand Russell tells a story something like the following:- A priest, centuries ago, was asked to observe the number of planets visible through a recent invention, a telescope. He refused, stating there is no need to observe in order to discover the truth, because the Bible says there are seven (a holy or special number, according to the Bible). The telescope could show there were at least eight.

[88] A reminder: ‘UCI’ is short for the ‘universalisability version of Kant’s alleged categorical imp