This book is about the idea of action without desire in Indian philosophy – a yogic idea paradigmatically expounded and recommended by Kṛṣṇa Vāsudeva in the Bhagavadgītā as that incarnate Lord sings, eventually successfully, to persuade Arjuna Pāṇḍava to kill his relatives and teachers in the Mahābhārata war. This idea theoretically allows the attainment of mokṣa (the end of a karmic series of lives) without the need for renunciation of one's societal duties; and it is discussed in many surviving ancient and medieval Indian texts – often, as by the great Vedāntins Śaṅkara and Rāmānuja, in commentaries to the Bhagavadgītā itself.

Christopher Framarin's book delineates the idea in philosophical and theoretical terms, wondering, from a variety of well-painted and well-linked angles, how a desireless yogī might yet act. Integrating material from several of his recently published papers, Framarin experiments with many lines of possible explanation, presents them in formal terms with numbered premises and conclusions, and picks holes in them; and at the end he remarks upon what remains. He begins with the Bhagavadgītā and refers to it frequently, bringing other textual material to bear on the discussion as and when it suits his methodical progress.

After a preface and brief introduction, chapter 1 sets aside the ‘absurdity interpretation’ (whereby Kṛṣṇa must be talking nonsense), the ‘no action interpretation’ (whereby what the yogī does is not really action, so the problem of action without desire does not arise), and comes to the ‘some desires interpretation’, whereby the yogī is free of certain desires or types of desire, but still has whatever sense of ‘desire’ might be necessary functionally to produce action.

Exploration of the some-desires interpretation then occupies most of the rest of the book. This interpretation takes the doctrine of desireless action non-literally, qualifying Kṛṣṇa's apparent instructions to Arjuna. Such a non-literal interpretation would be necessary if ‘actions’ were always desideratively linked to specific agents: ‘David Hume begins his discussion of the topic by admitting … his view, according to which desire is entailed by action’ (16). Tolerating the Humean position for now, Framarin tries to work out what kind/s of desire would be permissible for a yogī. Chapter 2 sets up the topic of permissibility in desire, referring to the difference and overlap between ends and means, and to Patañjali's Yogasūtra and Vyāsa's commentary thereon. It discusses the perfectly knowledgeable agent – for example, Kṛṣṇa in the Bhagavadgītā, and/or Īśvara in the Yogasūtra, both of whom are desireless yet motivated; and hypothetical yogīs, always knowing the difference between the true self and the non-self, are of the same stamp – and argues that permissible desires must be ones that a perfectly knowledgeable agent could have.

The following chapters explore three versions of the some-desires interpretation, that is, three proposed criteria of permissability for yogic desires: as desires for mokṣa; as unselfish desires; and – with reference to the Nyāyasūtra and the Brahmasiddhi and their commentaries – as phenomenologically non-salient desires. In each case objections are considered and adjustments made, but the proposal fails: these lines of thought do not yield any coherent criterion of permissibility. So Framarin concludes that the ‘some desires interpretation’ fails. In the final chapter, ‘A defense of desireless action’, Framarin reverts to a literal understanding of Kṛṣṇa's idea. This can only be done by tackling the Humean view according to which action without desire would be impossible. So Framarin considers recent analytic work on desire in action, and refutes the Humean view.

One might suggest that had he done this earlier, when he first introduced the some-desires interpretation, then perhaps it would not have been necessary to take that interpretation quite so seriously, and some of the lucubrations of the intervening chapters might have seemed less pressing. But the depth in which Framarin interrogates the eventually unhelpful ‘some desires interpretation’ is justified by the prevalence of varied statements of that interpretation in modern commentaries, and as such his book is both welcome and important. He convincingly defends the notion of literally desireless action against clear and present lines of philosophical attack. His desireless yogī can act, because whatever inherent aspect of ‘action’ it was that the Humeans thought was necessarily to do with desire is not so – but is no less inherent for all that. In defending desireless action, Framarin shows that desire is not the key concept in understanding human action, thus returning us to Kṛṣṇa's suggestion that the best actions – the paradigmatic actions, such as those of King Janaka – cannot be positively analysed in terms of desires.

The strength of the book is its neti neti method (it can't be like that, and it can't be like that either). Various ways of stating proposed ideas are held up to the light, adjusted, found to be opaque, and discarded. One might only wonder what satisfying statement could pass the test; Framarin leaves those who would explain Kṛṣṇa's idea much better off, but it is not clear whether he does so himself. Perhaps no-one can; but in what follows I criticize some aspects of the book, and suggest further lines of approach.

Where Framarin discusses Indian materials other than the Bhagavadgītā, one of his principal concerns is to show that the Humean view – of which the ‘absurdity interpretation’, set aside as uncharitable in chapter 1, is an extreme version – was not shared across the Sanskrit tradition. He wants to show this in support of his overall argument for taking Kṛṣṇa's idea literally. But in chapter 5 one might question Framarin's contention that the Bhagavadgītā and the Manusmṛti present similar theories of motivation. Kṛṣṇa, taken at face value, does not share the Humean view; but Manusmṛti states that ‘[T]here is no desirelessness in this world … [N]ever is any action of a desireless one seen in this world’ (na caivehāsty akāmatā … akāmasya kriyā kācid dṛśyate neha karhicit, 2:2–4), which could be a Humean denial of Kṛṣṇa's theory (even if Manusmṛti 12:88–90, which Framarin does not mention, presents a more compatible view).

Framarin has recourse here to Medhātithi's commentary, and then projects it back onto the Manusmṛti in a way that I find unconvincing. Speaking generally, the contextual distance between the commentators and the root texts is potentially extreme, and using the commentaries in an attempt better to understand those texts is a risky method, often pursued at the expense of those texts. Some readers may greet a book on desire and motivation in Indian philosophy with more zest than they would a book on desire and motivation in any single old Indian text; but I am not among them, for all that what the commentators say is nonetheless fascinating. I suspect that desireless action in the Indian context is best explained in the Bhagavadgītā, and in the Mahābhārata of which it forms part. Both of those texts would be a more convincing locative object of study than ‘Indian philosophy’, especially as this is curtailed by Framarin's use only of brahmanical sources.

The projection of the commentarial voice onto the text is clear also in chapter 1, when Framarin sets aside the ‘no action interpretation’ of Kṛṣṇa's words. Discussing the Bhagavadgītā's deterministic account of human action, he seems to feel that removing intentionality from human behaviour would disqualify it as ‘action’. Here he writes in the wake of a good deal of philosophical work on human action in recent decades, much of which has been, I believe, overdetermined by systematic academic service to a societal framework of statable moral responsibility. Śaṅkara and Rāmānuja too, introduced by Framarin on 12–13 to make his point, are concerned to leave room for the individual to make a motive difference between doing and not doing the right thing. But for me this has less to do with the facts of action than with human traditions of judging individuals.

Framarin projects the denial of determinism onto the text while sidestepping Bhagavadgītā 11:32–4 and 18:59–61, where Kṛṣṇa says Arjuna will fight whether he likes it or not. But viewing the idea of desireless action in context, one might suspect that the absence of free, judgeable will was a vital aspect of it. So we ought at least try to deconstruct that will; cf. Daniel M. Wegner The Illusion of Conscious Will (Cambridge MA: MIT Press, 2002).

There is indeed a narrative difference between some human behaviours and others, upon which culpability is erected. In a counterfactual narrative, an actor might have done otherwise; and in the narrative of conventional responsibility desire is an aberrant force, to be conquered and controlled. The dangers of the king's unchecked desire are widely illustrated in the lore, and the scapegoating mechanism (whereby others are condemned and punished in the good king's name) works by suggesting that if it had been me, I would have had more self-control – I would have been able not to do that. Well might a text recommend the eradication of desire from royal action, even to Arjuna who will never be king.

But Kṛṣṇa is also saying, more seriously, that the imaginary concentration of responsibility upon a formal individual is ahaṃkāra, the reification of an ‘I’, and is to be discouraged, for it is thinking about how deeds might reflect upon a false self – inevitably also one's own false self – and so causes suffering and karmic bondage. Hence the attempt freely to do the right thing (or act in the right manner) is a red herring. Arjuna fights not because he is persuaded to, but because he is powerless not to. Yet he is not forced to, for the knowledge that he is powerless not to reverses the value of any operative distinction between the him-that-does and the it-that-is-done. His mind changes when it is revealed as not his. And yet he is urged to train his own mind!

Framarin's espousal of the commentatorial position involves regarding Kṛṣṇa's exhortations to Arjuna – to get up and fight – as evidence that it is up to Arjuna whether or not to fight (‘Śaṅkara and Rāmānuja point out that … the Gītā … enjoins certain actions, which implies freedom’, 13). But this, I suspect, is a fallacy. Rather, one might suppose that exhortation is common because being told to do something can make one do it (as with Yudhiṣṭhira and the Pāṇḍava rājasūya); and this can seriously discourage the idea that when one acts or does not act, one is freely choosing. Perhaps the more Kṛṣṇa (his maternal cousin, brother-in-law, and blood brother) exhorts Arjuna to fight, the less sense it makes to think of Arjuna as culpably responsible for fighting. In the Mahābhārata, because Kṛṣṇa made Arjuna fight, it is Kṛṣṇa and his people, not Arjuna and his, who are cursed by Gāndhārī to suffer the consequences that Arjuna initially fears from this deed. And Kṛṣṇa's exhortations would already be just as constrained as Arjuna's fighting will be. The Mahābhārata invokes this context too, by explaining about Kṛṣṇa's unchosen celestial mission to make the war happen and so to effect the deaths of cosmic miscreants embodied in human form.

So when Framarin dismisses what he calls the ‘no action interpretation’, the label is questionable, and may rest on a translation problem. The proponents of this interpretation would not say that Kṛṣṇa suggests not performing karman (the word normally translated ‘action’); its characterization as a ‘no action interpretation’ (and as a ‘non-literal interpretation’) depends on a temporary adoption of the same kind of Humean position that Framarin eventually rejects. So it is not clear how much of the argument against this kind of interpretation would remain at the end of the piece.

Framarin fails to differentiate my published views (see ‘Calling Krsna's bluff: non-attached action in the Bhagavadgītā, Journal of Indian Philosophy, 32 (2004), 81–103) from those of George Teschner. Framarin writes (9):

That may be Teschner's view; but it is not mine. The first part is correct, but the final move, which Framarin ascribes to me repeatedly on 9–11, is wrong; rather, many behaviours are intentional, and those are the ones liable to leave a karmic residue and condition rebirth. My impression is that although many actions are conditioned by intentions, the yogī's are not – even if observers might imagine that yogī intending lokasaṃgraha (the holding-together of the worlds), or the wholesome good of sentient beings, or some other lofty formal goal.

Much depends here upon the word saṃkalpa, which denotes a motival aspect commonly allied with desire, and is often translated as ‘intention’ or ‘purpose’. Framarin points out that Manusmṛti 2:3 presents saṃkalpa as the root of desire. This formulation is paralleled at Bhagavadgītā 6:24 and Mahābhārata 12.171:25. Thus saṃkalpa would seem to be unlike, for example, kratu at Bṛhadāraṇyaka Upaniṣad 4.4:5, which results from desire and conditions action. Framarin's sketched position leaves room for something like saṃkalpa to motivate the yogī's action without generating desire. He surprisingly suggests, on the basis of Medhātithi's commentary, that saṃkalpa at Manusmṛti 2:3 means ‘belief’; but when the word occurs in the Bhagavadgītā Arjuna is advised to eradicate it in the same way, and for the same reasons, as he is advised to eradicate desire. At Bhagavadgītā 4:19, 6:2, and 6:4 it is clear that the yogic actor has renounced all saṃkalpas; and one immediate implication of the ‘root’ analogy might be that while saṃkalpa survives, desire will grow back. At the end of Framarin's book, one can imagine a set of enquiries about non-literal interpretations of Kṛṣṇa's advice to eradicate saṃkalpa; and perhaps this is the next step.

Some formal points to finish. The conclusion ends on 126; 127–171 present three appendixes – translations of and commentaries on commentaries – which supplement the arguments of chapters 5 and 6. These are perhaps not as well articulated with those chapters as they might have been – I missed the signal referring the reader to Appendix II – but they are nonetheless of independent interest. The index needs more sub-entries. The transliteration conventions used in presenting Sanskrit text are idiosyncratic in not separating words that would be unseparable only in other scripts. The positioning of the notes (including Sanskrit text) as endnotes rather than footnotes is a nuisance, and the book is badly overpriced as a physical object, but it bespeaks skilful, original, and valuable work, in an important and fertile field.