Bryn Mawr Classical Review 96.10.14


Dominic Scott, Recollection and Experience: Plato's theory of learning and its successors. Cambridge: Cambridge University Press, 1995. Pp. ix, 289. $54.95. ISBN 0-521-47455-8.


Reviewed by Allan Silverman, The Ohio State University

The expanded title of this work is a better indicator of its concern and range: Plato's theory of learning and its successors. Scott is the first to admit that he does not aim to explore all that one might consider learning in Plato. Nor does he undertake the impossible task of tracking the history of the notions of recollection and experience from Plato to the present. The discussion of Plato is roughly a third of the book, and one of Scott's goals is to defend the interpretation of recollection developed in Scott, D., 'Platonic Anamnesis Revisited.' Classical Quarterly (1987): 346-66. The second part analyzes Aristotle's empiricist response to Plato. The remainder examines Epicurean and Stoic accounts of learning, with the focus on showing that their concern with 'ordinary learning' and dispositional innatism is a departure from Plato and Aristotle and the source of the Post-Renaissance doctrine of innate ideas. The last two chapters are brief sketches of the innatism of the Cambridge Platonists and the curious case of John Locke. In a brief conclusion, Scott traces how the innatist/empiricist debate plays itself out in contemporary moral philosophy. For the purpose of this review, I will say little about these modern periods.

Plato's theory of recollection and its place in the history of innate ideas is an exceptionally worthwhile subject. Much of what Scott says is controversial, and a careful reading will benefit anyone interested in the topics of innateness or ancient theories of learning. Nonetheless, while Scott concedes that a modest encyclopedia would be needed to do justice to all the issues involved in a study of learning over such a span of years, this study is too brief and cursory. None of the successors enjoys the treatment he deserves. The Hellenistic authors receive surprisingly short shrift -- 50 pages. The discussion of Aristotle seems more an adumbration than an apodeixis, as Scott tries to incorporate into his argument supremely difficult sections of the Posterior Analytics, De Anima, Metaphysics, Physics and Ethics. And since there is little in the chapter on Plato that is not to be found in the CQ article, critical elements of the Platonic account are left unexplored. The book is more a tantalizing morsel than a full meal.

The focus of the book is 'three issues to do with learning and discovery.' (3) The first is the distinction between 'innatist and empiricist theories of learning... [T]his is a distinction between theories that appeal to the mind's internal resources to explain many of its cognitive achievements and those that explain them from the external input of sense perception.' (4) The second focus 'concerns the level of learning to be explained.... All the ancient philosophers included in this study made some distinction between different levels of thinking: the philosophical or technical and the mundane or pre-philosophical.' (5) A theory of learning will thus have to explain two different processes, namely how we begin or how we acquire concepts and language, and how we advance from this ordinary learning to the more sophisticated, technical or philosophical knowledge. Given that there are these two levels or perspectives, the third focus concerns 'the relationship or, better, the distance, between these two perspectives.' One can maintain that the gap is unbridgeable, -- to advance to the higher level one must 'reject' the ordinary beliefs -- or bridgeable -- one perhaps need only refine ordinary learning. Depending on how one views the 'gap', one is liable to be a pessimist or an optimist about ordinary learning and consequently about the likelihood that most or many of us will achieve the higher levels of learning, and whatever comes from that achievement.

Plato's theory of recollection, according to Scott, is designed to explain not ordinary concept acquisition but philosophical knowledge. His Demaratan account, named after the Greek double agent who hid a message under wax, assigns to sense -- perception and the give and take of conversation our ordinary learning: Plato is an empiricist when it comes to the acquisition of ordinary concepts. Recollection thus does not contribute to our ability to classify particulars or range them under concepts -- the Kantian reading. It is designed to account only for knowledge of Forms. Recollection begins with the process of philosophizing and hence only philosophers, or those on their way to becoming such, recollect. There is no commonality between the recollected concepts and the ordinary concepts save for the fact that sensible particulars have a share in Forms. Plato then is a severe critic of common sense both because it is false and also because it cannot be utilized in the philosophical process to achieve knowledge. The gap is unbridgeable. Hence Scott's Plato is unqualifiedly pessimistic about the majority of us who rely on common sense.

Aristotle rejects Platonic recollection but this does not qualify him as a tabula rasa empiricist. Innatism has at least two non-Platonic varieties, on Scott's reckoning. Whereas Platonic recollection requires that one at some time be aware of the concepts to be recollected, an alternative maintains that these concepts are merely latent, i.e., awaiting some stimulation but not requiring that one be aware of them at some moment prior to their arousal through sensory stimulation. A second, more amorphous form of innatism dispenses with concepts in favor of dispositions to form either concepts or beliefs. The structure of the mind or human nature is such that we are disposed to form beliefs or concepts regardless of how we are raised or what we confront in the environment. Aristotle is, according to Scott, tempted by dispositional innatism in moral philosophy, since we are inclined to behave in certain ways by our very nature. But even here experience turns out to be critical. The behavior towards which we are inclined plays no role, says Scott, in moral education and hence his Aristotle is not an innatist of any sort. Following upon this discussion Scott argues that the notorious final chapter of the Posterior Analytics is not about ordinary concept formation but rather scientific or philosophical knowledge. Unlike Plato, however, Aristotle thinks that experience is to be ranged on a continuum such that through refinement and articulation we can turn the fuzzy and partial knowledge we derive from sense-perception into the philosophical knowledge of definitions and first principles. This is the process by which we move from what is knowable by us to what is knowable by nature. Because this is viewed by Aristotle as a continuum, prospects for the human condition are less bleak than those we found in Plato.

The chapters on the Hellenistic Schools aim to answer the question whether the invention of the theory of innate ideas can be dated as far back as Epicurus, or the early Stoa, or whether it was a later phenomenon. The notion of innate ideas at play here differs from the Platonic doctrine. The Hellenistic authors are concerned with ordinary learning and dispositional innatism. What is important for understanding the emergence of the 17th c. doctrine is to test whether one thinks that the beliefs or ideas formed from the innate dispositions are 'epistemologically basic, acting as the foundation of moral, theological or logical knowledge.' (188) The accounts of Epicurus and the Stoics are fertile ground for innatists, it seems, since both deploy 'prolepses' and 'common notions' quite liberally throughout their philosophy. Beginning with Epicurus, Scott contends that, despite his reliance on preconceptions as a criterion of truth, Epicurus is not an innatist. Epicurus would not admit that sense-experience is inconsequential with respect to what concepts and beliefs we develop. Scott claims that only in the case of the gods are there grounds to suspect innatism in Epicurus. But he argues against Long and Sedley's innatist account. The crux of the argument is how to understand Cicero's remark at N. D. I. 44, quoniam insitas eorum [sc. deorum] vel potius innatas cognitiones habemus. Scott argues, as others have before him, that innatas here refers to the natural way in which the ideas arise, as opposed to a conventional account of the origin of the idea. So there is no case of innate ideas in Epicurus.

Scott contends that the Stoa is the source of the modern notion of innate ideas, but in a limited fashion: only with respect to moral notions and beliefs are they innatists. According to his account, the early Stoics maintain that we are disposed to behave virtuously. We then need an argument to get from our disposition to behave in a certain way to the formation of the appropriate concepts. This is found in a theory of behavior according to which 'action presupposes beliefs and notions, so that if we are innately disposed to act in certain ways then we must be innately disposed to form the beliefs and notions necessary for that action.' (206) Such a theoretical bridge is provided, says Scott, in the Stoic notion of assent. Assenting to a bit of behavior requires the notion of the good. This and other moral notions turn out to be innate, though sense-experience will play a crucial role in developing them. In contrast are notions such as black and white. While it might seem that they too are part of our nature, the role of sensory experience is not just crucial but direct -- we directly experience black and white. Apparently then either we do not directly experience moral notions, or the direct experience is not necessary for their development.

It is difficult to see how we are to derive the thesis that these Epicurean or Stoic notions are epistemically basic in any sense. Scott's discussion of the Epicurean critique of common sense reveals that there is nothing about the natural prolepses that allows one 'subjective clarity', i.e., the power to know that this prolepsis is natural, or that one not. But if subjective clarity is not forthcoming, how are the Stoic or Epicurean ideas the source of the subjectively clear modern notion of an innate idea? On the other hand, the different attitude towards common sense on the Stoic side does forge a connection with the modern position. If something is innate it should be common. And if common, it should show up in all of us and in our actions and beliefs, i.e., accounts of our actions. Epicurus attacked common sense because he thought that (false) beliefs were super-added to our sense-perceptions and prolepses. The Stoics attack every man's beliefs in a less virulent form, and they are less pessimistic about our prospects. Stoics maintain that moral beliefs are trustworthy because they derive from innate dispositions. And the reason why this makes them trustworthy is that nature is 'providential', i.e., good. This does have the ring of a theological defense of innate ideas, whose presence is palpable in post-Renaissance discussions.

It is worth noting two themes from the modern period that configure Scott's treatments of the ancient material. The first is how and when we are 'aware' of innate ideas, concepts and beliefs. Certainly 'awareness', or whatever notion is at play in the Cartesian account of the cogito, is a critical concern for the moderns. But it is no easy matter to botanize the importantly different kinds of epistemic access to our own mental states. For that reason alone it is hard to know what to make of Scott's efforts to distinguish the different ancient accounts in terms of when, how and whether we are aware of an innate idea. For instance, Scott spends little time on the possibility that Platonic recollection, or the 'innate' Forms, enable embodied souls to sort out the mass of incoming information into neat, or somewhat neat bundles, and that we do so without (ever?) being aware of their performing this function. Hence, his account leaves it rather mysterious how we come to have, for the most part, ordinary concepts for which there also happen to be Forms. While it is true enough that Forms provide the metaphysical basis for the similarity, Plato is aware that we could be systematically and globally wrong about what we experience. (Witness the Heraclitean account rejected in the Cratylus.) If ordinary experience is self-contained and internally coherent, and the innate Forms do nothing until the philosophical process commences, then there is no reason for that process to commence. For we note the 'contradictions in our sense-experience,' not from the perspective of sense-experience, but from the philosophical perspective. Hence on Scott's Demaratan account we should never note the contradictions. Plato, I submit, thinks that we do not adopt a coherent but false account because our language and beliefs are developed in part through the workings of the innate ideas, the once-viewed Forms, even though we are not aware of it happening. (Of course after philosophizing we are aware that such a process must be occurring, since we cannot otherwise explain our abilities.)

A second and, I think, more startling omission from Scott's account of the modern and ancient debate is the topic of simplicity. For Descartes, the criterion for the epistemically basic is the simplicity of the innate idea. (Cf. Meditation V) When we turn back to the Platonic and Aristotelian accounts, we find that the simplicity or unity of the definition is also critical. Strictly speaking, primary substance is simple. This simplicity, along with the visual metaphors, is I think also at the root of the Platonic account of the unity and simplicity of the Form and provides a reason why Plato repeats the visual metaphors throughout all of his accounts of how we know the Forms. The heart of intuition, or intuitionism, if you will, is the gap between the intuitive insight into the unitary nature of the Form and the syntactic and semantic complexity of a verbal account, i.e., linguistic definition, of what we know.

Finally, Scott's account of the Stoic origin of the modern idea is quite a stretch. Even if we accept that the Stoics thought that we have innate dispositions to behave virtuously, the texts allow us to claim only that they should have developed a corresponding account of innate moral notions. If we believe strongly in their perspicacity, we can say that they did, since the texts are lacking. Moreover, given the theory of assent, Scott's argument would go through without any dispositions to virtuous behavior. Provided that we are disposed to act, and action requires assent to the proposition that 'it is (seems) good to do x', then by Scott's lights we should be disposed to develop the notion of the 'good' -- or the seeming good?. But this seems to say little more than that we are by nature rational animals who act. It is a far cry from the modern arguments on behalf of innate ideas.

I have remarked that the breadth of the discussion is quite great. It will behoove anyone interested in the notion of innateness to look at this book. There are some excellent discussions, especially of Plato, which repay the attention of the reader. But for all its breadth, the discussion is not very deep. Innateness is much discussed, especially in the contemporary linguistics literature. Scott makes little use of this material. I would urge the Classicist interested in the continuation of the debate to read Chomsky.