Russell Contra Sense/Reference, the ‘Mont Blanc’ Correspondence

ABSTRACT It is argued that Russell before 1905 saw no value in Frege's sense/reference distinction. This is clearest in the Mont Blanc correspondence. It is argued that Russell and Frege failed to engage because Frege lacked a grasp on the internal/external relations distinction. For Russell sense is either an external relation, objectionably separating out thought and reference, or an internal relation, so what is thought is altered such that we do not know what we are talking about. The novelty of the present paper lies in the arrangement of the parts and the claim that Russellian propositions are not made up of the things themselves but of transparent representatives thereof.


Introduction
Russell discusses Frege's sense/reference distinction in Appendix A of Russell 1903, commenting thus ( §476): The distinction between sense (Sinn) and reference (Bedeutung) is roughly [. . .] equivalent to my distinction between a concept as such and what the concept denotes ( §56) [. . .] the reference of a proper name is the object which it refers to; the presentation which goes with it is quite subjective; between the two lies the sense, which is not subjective and yet is not the object. 1 The passage Russell alludes to is this (Frege 1892, p. 60): The reference of a proper name is the object itself which we designate by using it; the idea which we have in that case is wholly subjective; in between lies the sense, which is indeed no longer subjective like the idea, but is yet not the object itself.
That Russell regards Frege's sense/reference distinction as 'roughly, though not exactly, equivalent' to his 1903 theory of denoting, is liable to mislead his readers.This will become clearer in due course.For the moment, whereas sense/reference is general, denoting applies only to concepts preceded by a syncategorematical term, one of all, every, any, a, some, or the (Russell 1903,  §58).Denoting is intended to ensure an entity referred to in such cases. 2It follows, according to Russell, that denoting is indefinable, a 'fundamental notion of logic' ( §56).Unlike sense/reference it is not a cognitive notion, because at this date Russell's anti-psychologism is at a peak and, it will be argued, he sees Frege's sense/reference distinction as cognitive and not logical and therefore (by default) as psychologistic. 3As should become clear, part of the reason for Russell and Frege arguing past one another is that they effectively have a different view on what is, and what is not, psychologistic.
Sticking with proper names, Frege proposes a tripartite account; (i) the wholly subjective and possibly individually idiosyncratic idea, (ii) sense, intended as a wholly objective and publicly available mode of presentation of or route from the idea to the reference, and (iii) reference, the thing itself.The account is intended to generalise to the constituents of thoughts generally, including definite descriptions (the Evening Star, the centre of the earth, and so on), and numerals ('7').
It is evident in Russell 1903, §476 that Russell has no use for sense, which he sees as placed awkwardly between ideas and the things they are of.This comes to a head in a 1904 exchange of letters between Frege and Russell.The central contention of the present paper is that the two argued past each other, because they approached the problem with radically different philosophical repertoires.Consequently in the Mont Blanc correspondence they failed to get to the heart of the matter.To put the matter more sharply, it is not the case that Russell rejected a clear-cut Fregean distinction.Rather, Russell couldn't see any use for a distinction that, to him, awkwardly straddles the boundary between the logical and the psychological.
The aim in this paper is to throw light on Russell's early work, prior to Russell 1905, and on the pressures that led to the theory of descriptions contained in that paper.It is not argued that Russell had a way of addressing Evans's intuitive criterion of difference, 'that the thought associated with one sentence S as its sense must be different from the thought associated with another sentence S as its sense, if it is possible for someone to understand both sentences at a given time while coherently taking different attitudes towards them, i.e. accepting (rejecting) one while rejecting (accepting), or being agnostic about, the other' (Evans 1982, pp.18-19).The cognitive shortcomings of Russell's pre-1905 work stand in need of acknowledgement.Equally it is not argued that the views advanced assist with Russell's difficulties with the unity of the proposition.
The specific issues discussed here remain opaque and have seen little consideration, in part because of a widely held view that Russellian propositions are made up of the things themselves.Hylton quotes from the Mont Blanc correspondence at Hylton 1990, p. 172, remarking that 'Propositions and terms are thus notions with immediate ontological significance '.It is striking how often the claim that Russellian propositions are made up of the things themselves is justified not on the basis of Russell's published writings but on the basis of one passage in a 1904 letter to Frege (quoted below). 4 This view is, I think, mistaken, because it can't explain how we can think about and talk about what there is.

'Mont Blanc'
Mont Blanc makes its appearance in Frege's letter to Russell of 13 November 1904, in the context of remarks on truth (Frege 1980, p. 163): I agree with you that 'true' is not a predicate like 'green'.For at bottom, the proposition 'It is true that 2 + 3 = 5' says no more than the proposition '2 + 3 = 5'.Truth is not a component part of the thought, just as Mont Blanc with its snowfields is not itself a component part of the thought that Mont Blanc is more than 4000 m high.
Later in this letter Frege turns to sense/reference (Frege 1980, p. 164): We thus find that the thought depends on something other than what is referred to by the sign; for this is the same for '7' and for '4 + 3'.A sign must therefore be connected with something other than its reference, something that can be different for signs with the same reference.Signs do not just refer to something; they also express something.This is the sense.
Russell replies to this in a letter of 12 December 1904.First, he takes Mont Blanc, sense and reference, and truth, together (Frege 1980, p. 169): Concerning sense and reference, I see nothing but difficulties which I cannot overcome [. . .] I believe that in spite of all its snowfields Mont Blanc itself is a component part of what is actually asserted in the proposition 'Mont Blanc is more than 4000 m high '.We do not assert the thought, for this is a private psychological matter: we assert the object of the thought, and this is, to my mind, a certain complex (an objective proposition, one might say) in which Mont Blanc is itself a component part.If we do not admit this, then we get the conclusion that we know nothing at all about Mont Blanc [itself 5 ].This is why for me the reference of the proposition is not the true, but a certain complex which (in the given case) is true.
So in these passages we have views on truth, sense and reference, signs (numerals) and numbers, Fregean thoughts, and propositions and their constituents.Less obviously we have views on relations, because relations obtain between ideas, senses, thoughts, and references, as well as between signs (numerals) and numbers.The locus of Russell's and Frege's inability to understand one another centres on the issue of relations that do or do not obtain between these.

Sense and Reference
For Frege, the sense/reference distinction is most obviously required to account for the informativeness of identity statements, cases where names/definite descriptions refer to the same object.If one doesn't know that the Evening Star is the Morning Star, being told that the phrases are coreferential augments one's knowledge.If the constituents of a thought are the things themselves, this cannot occur.It should be immediately evident that they are the same, and hence that any true statement of identity wears its truth on its face.Since this is not the case, the constituents of a thought cannot be the things themselves.
Frege does not accept that the subjective idea(s) one associates with a name or a definite description will do, as such ideas are insufficiently objective for logic.So he needs something to fulfil a particular theoretical role.This is where sense comes in, performing an explanatory role as accounting for the (potential) informativeness of statements of identity, when one doesn't know that co-referring terms are so.But motivating this on the grounds of the informativeness of identity statements is perhaps too close to psychologism. 6It is arguable that Frege states the sense/reference distinction as consequent on the informativeness of statements of identity as an expository or elucidatory strategy, but this is rather weak when what is wanted is a justification on purely logical/objective grounds, with no taint of psychologism.
Dummett holds that Fregean sense is a cognitive notion; 'The notion of sense is introduced in connection with that of knowledge: it is required in order that we may give an account of how we know the references of the expressions of our language' (Dummett 1981,  p. 229; cf.p. 240).But there remains a very real difficulty.If senses are wholly objective third realm entities, then there is no obvious way of accounting for our knowledge of them.This is a familiar problem, that there is no obvious way in which we can get to know what is acausal and non spatio-temporal.Frege's notion of grasping is a metaphor, a way of stating a problem and not a solution.By contrast Russell's 1903 theory of denoting is wholly logical.What it does not do is offer any non-accidental account of cognitive value, that is, how we know what the object is that is denoted by a denoting concept.
If senses as objective entities fall entirely within the domain of the logical then we have no handle on how we get to know them, on their cognitive value.But they are not supposed to be subjective, psychologistic entities either.This is a difficult tightrope to walk, and this is why Dummett bristled at Sluga's assertion that he had pursued 'a conflation [of semantics and epistemology, in a Fregean context] further than anyone else' (Dummett 1981, p. 530).
At stake here is what one takes to fall within the domain of the logical, and what one takes to fall within the domain of the psychological.This is not clear-cut, because it is not absolutely clear where the boundary lies.In fact Frege is in the process of establishing a demarcation, with the sense/reference distinction sitting rather awkwardly astride.Whether one can have a cognitive notion of sense that is not in any way psychologistic is a moot point. 7n Frege 1879, §8, pre-dating the sense/reference distinction, Frege asserts that in the case of statements of identity, what flanks the sign of identity is names, and not contents.This is in contrast to other forms of judgements, where we deal directly with contents, with the things themselves.In Frege 1879, different names for the same content are associated with a mode of determination (Bestimmungsweisen).Frege does not state how modes of determination pick out or identify a content.We are told that a judgement of identity is synthetic, but not how we perform such a synthesis.Implicitly, in Russellian terms, the account given in Frege 1879 is purely logical.
That a Begriffschrift mode of determination has effectively the same logical characteristics as a 1903 Russellian denoting concept goes together with an inability to account for informativeness, that is, cognitive significance.This led Frege to introduce the sense/reference distinction in Frege 1892, the intention being that the distinction is both cognitive and logical.But we are still not told how senses as modes of presentation or routes to references are supposed to do what they do.As Evans points out, 'Frege never said much about particular ways of thinking of objects; he provided no analysis of what it is to think of an object demonstratively, for example' (Evans 1982, p. 18).
This brings into sharp focus Frege's phrase 'cognitive value' to account for the difference between a = a and a = b (at Frege 1892, p. 56), that if we take a = b to mean that a and b are both names for the same content, then if we take a = b is true as a statement that a and b both name the same content 'we would express no proper knowledge by its means' (p.57). 8e would merely have said something about names, that they are taken to be coreferential.Introducing senses as objective, as entities that are supposed to fall within the domain of the logical, is intended to ensure that in dealing with statements of identity one remains within the logical, and doesn't lapse into the psychological.This pulls questions concerning knowledge (implicitly taken to be factive) within logic.If Frege had filled this lacuna, then Russell should have grasped the difficulties involved.Sense would be a settled affair, not a bone of contention.But as Dummett points out, 'even when Frege is purporting to give the sense of a word or symbol, what he actually states is what its reference is' (Dummett 1981, p.  227).Evans notes that 'Frege had no more idea of how to complete a clause like 'The sense of "and" is . . .' than we do' (Evans 1982, p. 26).The elephant in the room is that any word in the English language can be intelligibly substituted for and.It should not be overlooked that Frege has effectively extended the domain of the logical, as compared to Frege 1879.It is this extension that Russell somewhat inarticulately responds to.

Idealism and Relations
Innocent of late nineteenth-century British idealism, and not obviously immersed in idealist philosophy generally, Frege is not oversensitive to, in particular, questions of relations. 9So the question of the relations between one's ideas, senses as modes of presentation/routes to references, and references, is not, it seems, of great concern.There is something that we do-refer to things-and there is some way that we do it, such that the requirements for logic are met.Beyond this lies psychology, outside the purview of logicians and mathematicians.It must have seemed clear to Frege that this is what we do; here is an explanation of what we do; what more could be wanted?By contrast Russell is steeped in late nineteenth-century British idealism, and is profoundly sensitive to issues around relations.We have, on Frege's account, three sets of entities-ideas, senses, and references-and thus two relations-ideas to senses, and senses to references.What is the nature of the relation in each case, and how does it operate?For Frege these are largely non-questions; this just is what we do.For Russell this is not tenable.
The first question about relations is whether or not they are real. 10Assume that two entities are related, setting aside for the moment the nature of the entities (things, subjects, predicates, and so on).Is the relation a 'third thing', ontologically on all fours with its relata?If so it is hard to see how it can function as a relation, how it can couple or connect its terms.All remains separate.On the other hand if a relation is no sort of thing at all then it seems to evaporate and equally fail to do any relating.Given that our entities are related, then they appear to have collapsed together or perhaps into one another without any call for a relation. 11f we consider the Fregean theory, of idea-sense-reference, neither of these alternatives is palatable.Given that senses are supposed to do the relating, between what we have in mind and what there is, if sense is a 'third thing' then it stands aloof and while we may have a thought, it will not be of anything.If thought and what it is of collapse into one another, then we go straight from subjective idea to the reference, and we are back to a situation in which we cannot account for the potential informativeness of statements of identity.Whatever account of relations is to be given, we want a mid-path.We want sense to be not a something but not a nothing either, or at least, not something substantive.

Internal and External Relations
The alternatives sketched above are more familiar as external and internal relations.It isn't entirely clear when this terminology was introduced but by the turn of the century it was familiar within the philosophical milieu. 12Frege, however, shows no awareness of this distinction.There are many nuances to the doctrine, what matters here is how it was seen by Russell, and by Moore.The relevant aspects are set out in this and in the following section.
Russell's take is that an external relation is such that, should it obtain between two or more terms to a relation, the terms are unaltered by standing in the relation.However beyond asserting that they are related, no account is on offer as to how they are related. 13patial relations are paradigmatic external relations.Rearranging one's furniture changes the spatial relations between one's domestic chattels, but leaves the chattels themselves unaltered.
By contrast an internal relation is held to be such that standing in the relation in some way changes or alters the terms to the relation.This sounds straightforward, but is not easy to formulate.14Take, for example, a flatpack wardrobe.The components are the same whether they are in the box or constructed according to the instructions.It is only spatial orientation that has changed.So one might think that here, the relations are external, as all we have done is move things around relative to one another.
This overlooks the fact that something may be as it is because of the way other things are.One's wardrobe is as it is irrespective of one's chest of drawers, or at least, it may be the case that they are so independent (they may be from the same set or chosen to complement one another, but this is accidental).On the other hand a component of a flatpack is as it is because of the other components.Each is designed in such a way that they can be assembled into a whole, so their form is a joint endeavour.If a piece is missing one can work out how it should be (its parameters are determined) and engineer a replacement.If one's lodger absconds with the coffee table there is no equivalent (non-accidental) possibility.
The claim is that internal relations offer a ground or a reason or an explanation for the obtaining of relatedness, whereas external relations do not.If we think that things are internally related to one another, in extremis it follows that anything is as it is because of the way everything else is.Things are reciprocally determined.Such a view can be found in Bradley's Appearance and Reality.He begins by talking of what it is that two red-haired men have in common, going on to say this (Bradley 1893, p. 580): [I]f you could have a perfect relational knowledge of the world, you could go from the nature of red-hairedness to these other characters which qualify it, and you could from the nature of red-hairedness reconstruct all the red-haired men.In such perfect knowledge you could start internally from any one character in the universe, and you could from that pass to the rest [. . .] such knowledge is out of our reach, and it is perhaps out of the reach of any mind that has to think relationally.But if in the Absolute knowledge is perfected [. . .] the last show of externality has vanished.
That Bradley talks of knowledge is salient, because it is issues around knowledge that motivate the introduction of the sense/reference distinction.Whatever the metaphysics, Bradley's point is that if things are internally related, then in principle a complete grasp of any one thing entails a complete grasp of the universe.Of course we are not capable of this, because we have to think relationally.We can only think relationally because we are limited, fallible beings, incapable of grasping the Absolute, the totality of what there is. 15In similar fashion Frege brings in sense because there has to be some way in which we think about, identify and reidentify things; and bringing in the cognitive carries with it a risk of having to make non-logical concessions to our cognitive limitations.
At issue is the interaction between relatedness, and knowledge.The key point for Russell is that for something to stand in an internal relation to something else entails that it is changed somehow or altered by so standing.What needs to be borne in mind is the target-the nature of the cognitive relation, that is, the relation(s) idea-sense-reference, thought-world; everything said so far will converge on how these relations are to be handled.

The Axiom of Internal Relations
In discussing Joachim's The Nature of Truth (Joachim 1906), as representative of the Bradleian line of goods, Russell proposes an axiom of internal relations, that 'Every relation is grounded in the natures of the related terms' (Russell 1906, p. 139).
What lies behind this is an assertion that standing in a relation entails a change or alteration in the relata.This is what it is for a relation to be grounded.This is alluded to in Russell 1914 where, in considering a view urged by Bergson, Russell says (p. 157): This is part of a much more general doctrine, which holds that analysis always falsifies, because the parts of a complex whole are different, as combined in that whole, from what they would otherwise be.It is very difficult to state this doctrine in any form which has a precise meaning.
The point Russell seeks to make is that if the constituents of a complex are separated out they are different, in some not wholly straightforward sense, from how they are when they are constituents of a complex.Alteration or modification is perhaps more obvious in the case of vivisection than the dismantling of flatpack furniture.The change or alteration is of a rather different order of magnitude.
Russell goes on to say that 'the doctrine in question holds that a thing is so modified by its relations that it cannot be the same in one relation as in another' (Russell 1914, p. 157,  emphasis added).This is where Russell comes out in his true colours.Whatever the views actually held by Bradley or by Joachim, what matters here is Russell's interpretation of relatedness, in particular, of the difference between being internally and being externally related.For Russell an internal relation is one that ineluctably changes or alters or modifies its terms, whereas an external relation is such that something remains as it is whether or not it is related to something else.So whereas an idealist of the Bradley/Joachim school would say that something is as it is because of the way everything else is, and that to know anything fully requires a knowledge of everything else that there is, Russell's view is that if anything contracts into an internal relation it is thereby altered or modified in some as yet unspecified way.This is crucial to Russell's side of the Mont Blanc exchange.

Moore's Influence on Russell
Of Moore's influence, Russell remarks in the preface to Russell 1903 that 'In the more philosophical parts of the book I owe much to Mr G.E. Moore besides the general position which underlies the whole' (Russell 1903, p. xliv).The underlying philosophical position is stated in one of Moore's early essays (Moore 1904, p. 242, emphasis in the original): A good deal of confusion has, I think, arisen from the failure to see that the only alternative to the admission that we do know things as they are in themselves, is the admission that we have no knowledge at all.
In short, Moore's argument is that if we are to know what we are talking about, we must be in direct unmediated perceptual contact with it.The apparatus of forms of intuition internal and external, of pure concepts and categories, is entirely discarded in favour of an absolutely direct realism.It is only thus, Moore thinks, that we can be said to have knowledge, as opposed to a subjective or psychologically tainted opinion about what there is. 16evine points out that 'Russell holds that representations-specifically, linguistic representations-create no in-principle barrier to our accessing reality as it is in itself' (Levine 2019, p. 38), but he doesn't discuss the relations involved.Hylton points out that 'Moore's view of knowledge [. . .] is closely connected with that of the internality or externality of relations' (Hylton 1990, p. 125), and that the cognitive relation has to be external if we are to know what we are talking about (p.127), but this isn't in the context of the Russell/Frege correspondence.Further if, as Hylton does, one holds that Moorean/Russellian propositions are made up of the things themselves, that we have the things themselves in mind when we think propositionally, it is hard to see what the terms to a cognitive relation can be.17 Moore's views were first propounded thus (Moore 1899, p. 179): A proposition is composed not of words, nor yet of thoughts, but of concepts.Concepts are possible objects of thought; but that is no definition of them.It merely states that they may come into relation with a thinker; and in order that they may do anything, they must already be something.It is indifferent to their nature whether anybody thinks them or not.They are incapable of change; and the relation into which they enter with a knowing subject implies no action or reaction.
There is an immediate problem with the ontological status of Moorean concepts.The lacuna here is a direct analogue of the absence in Frege of an account of how senses are supposed to do what they do.Moore's intention is that what we have in mind when we think about what there is, is a product of direct and unmediated contact with the things themselves.For the moment this passage makes clear that Moore regards perceptual and cognitive relations between a knowing subject and what there is as paradigmatically external.Neither term to such a relation is in any way changed or altered or modified by standing in such a relation.18

Transparency
The Moore/Russell position requires rounding out with a theory of meaning, an explanation of how Moorean propositions as complexes of concepts, and Russellian propositions as essentially similar combinations of terms, get to be thought and expressed in sentences. 19ut given the direct realism in play it follows programmatically that little can be said, because any explanation that one tries to offer here interposes between the propositional as thought and/or said, and what there is.At this point there is a risk of losing sight of the fact (the claim, at least) that an external relation is intended to be a relation.Russell remarks that up to about 1917, 'I had thought of language as transparent-that is to say, as a medium which could be employed without paying attention to it '. 20Russell's idea is that one can, so to speak, speak what there is, rather than be ineluctably compelled to speak about what there is, whatever that might mean. 21urther, the doctrine of transparency underpins Russell's assertion that 'The correctness of our philosophical analysis of a proposition may therefore be usefully checked by the exercise of assigning the meaning of each word in the sentence expressing the proposition' (Russell 1903,  §46), that is, pairing off words and terms.Such an atomistic approach is only possible because propositional constituents (terms) are externally related to one another, thus ensuring that a term remains the same whichever proposition it is a constituent of (it is unaltered by the relations it stands in to other propositional constituents).

Miscommunications
That Russell has such views on the externality of relations in mind is clear in his 12 December 1904 letter to Frege, where he says that 'Mont Blanc is itself a constituent part of what is actually asserted in the proposition 'Mont Blanc is more than 4000 m high' [. . .] If we do not admit this, then we get the conclusion that we know nothing at all of Mont Blanc' (emphasis added).Here Russell echoes Moore's statement, in Moore 1904, quoted earlier, that either we know things as they are in themselves, or 'we have no knowledge at all'.But Russell doesn't appreciate that Frege is not steeped in late nineteenth-century British idealism, or the nuances of the Moore/Russell revolt against it.
That Frege and Russell talked past one another is shown by Russell's remark that 'concerning sense and [reference], I see nothing but difficulties which I cannot overcome '. (Russell to Frege 12 December 1904) 22 Russell doesn't say 'I think your views are incorrect, because . . .' or 'I think your views would be better stated . . .', and he doesn't propose an alternative theory.Rather he finds he is faced with a doctrine that does not compute, because their background positions and assumptions are too far apart.Whereas the young Russell steeped himself in the logics of Bradley, Bosanquet and Sigwart (as noted in the preface to Russell 1897), Frege regarded such works as 'thick logic books [. . .] bloated with unhealthy psychological lard, concealing all finer details' (Frege 1893, p. XXV).
The difficulty for Russell is this.There is what there is, there is what we have in mind when we think about what there is, and then there are the words-sentences-that we use to express this.This is nicely captured by Wittgenstein (Wittgenstein 1958,  §95,  §96): It is hard to think of a clearer statement of the early Moore/Russell position at the time Russell and Frege were in correspondence.This is alluded to by Moore (Moore 1903, p.  37): [I]n general, that which makes the sensation of blue a mental fact seems to escape us: it seems, if I may use a metaphor, to be transparent-we look through it and see nothing but the blue; we may be convinced that there is something but what it is no philosopher, I think, has yet clearly recognised.
The metaphor of transparency is intended to be more than a metaphor, it is intended to embody the notion of standing in line, each equivalent to each.As Hylton puts it, 'For Moore [. . .] the world is not alien to the mind but is, rather, transparent to the intellect' (Hylton 1990, p. 137).
Russell's difficulties with sense and reference now come into focus.When we think about Mont Blanc, what we have in mind is Mont Blanc, because it is only thus that we can be said to know what we are talking about. 23There is no additional mental fact that is the sense of or the mode of presentation of, there just is what there is, and not another thing.
However it does not follow that this can only be so if we have the things themselves in mind, because this is clearly absurd.What we have in mind is a suitably transparent, immediate representation or mental correlate of what it is we are thinking or talking about.As Hylton says, propositions are 'out there' (Hylton 1984, p. 14). 24But that they are mindindependent does not entail that they are constituted out of the things themselves.What is required is objectivity, not objectiveness.

Knowing What One is Talking About
For Russell, if a sense is a 'mode of presentation' or a 'route to a reference', then either it comprises an internal relation or an external relation between on the one hand proposition, thought, language, and on the other hand, what there is.It cannot be an internal relation because if it is, the terms are altered or modified, and thus we do not know what we are talking about.Mont Blanc as a term to an internal relation is modified so it is no longer Mont Blanc. 25n the other hand sense as an external relation leaves its terms unaltered, so knowledge is preserved.But now it doesn't do anything.It is marooned in no-man's land, neither the one nor the other, with no role to fulfil.This is why Russell cannot engage with Frege's position, because he simply has no use or need for sense as he thinks Frege presents it. 26iven their very different backgrounds and assumptions, drawing substantial conclusions from the Mont Blanc correspondence about the ontology of Russellian propositions is an exercise fraught with peril.
That a Russellian proposition as thought and as expressed in language is transparently representative is a tacit presupposition that allows Russell to conflate what is thought, what 23 Cf.Candlish 2007, p. 55. 24 Cf.Godden and Griffin 2009, p. 174. 25 Leading to a possibility of global reference failure, as Makin points out (Makin 2000, pp.153-4), with reference to Dummett; 'How can we know that we ever do reach the object, or that there really is any object, if a sense always interposes between us and it, a sense that carries no guarantee of any corresponding referent?' (Dummett 1981, p. 133 This is, I think, a fair representation of Russell's position prior to On Denoting, and I think if it had been put to him, he would have accepted it.Nevertheless the problem of the informativeness of statements of identity remains.The only straightforward way out is to shift the boundary between the logical and the psychological.This can be called an 'epistemological concession', a catering to the epistemological limitations of finite, fallible humans.Russell's problem with Frege's notion of sense is that it fails to do what is required, because while the intention is to shift the boundary of the logical with respect to the psychological, senses are either shifted into the domain of the psychological, and this is objectionably psychologistic, or sense is rendered logical, but Fregean senses so construed (as third realm entities) fail to convey us to the things themselves.This is, I think, the thinking behind McDowell's remark that 'it seems right to say the concept of sense belongs to psychology.But if we say this, we must conceive psychology otherwise than psychologistically' (McDowell 2005, p. 170). 28As far as Moore and Russell were concerned, though, circa 1898/1904, anything that even hints of psychology is psychologistic.
Russell's epistemological concession in 1905 is his doctrine of acquaintance, where acquaintance involves 'the things we have presentations of' (Russell 1905, p. 41), that is, 27 Evans says of Russell's position 'that the object in question actually occurs in the thought or proposition expressed by an utterance in which genuine reference is made [...] like a pea in a pod' (Evans 1982, p. 82).It is hard to see what one is supposed to make of this.There is always in the background Blackburn and Code's 'cardinal principle of Russellian exposition, to wit, don't make him out a complete fool if you can help it' (Blackburn and Code 1978, p. 67). 28As Godden and Griffin observe, Russell's shifts in his position (for present purposes between 1903 and 1905) are such that 'each of the changes he made to the theory of propositions moved him closer to psychologism' (Godden and Griffin 2009,  p. 175).
what is immediately given. 29What is more complex or mediate is known by description, and represented by a denoting phrase (Russell 1905, p. 56): Now such things as matter (in the sense in which matter occurs in physics) and the minds of other people are known to us only by denoting phrases, i.e. we are not acquainted with them, but we know them as what has such and such properties.
Russell says in a footnote that 'The theory [of denoting] there advocated [in Russell 1903,  Chap.V and  §476] is very nearly the same as Frege's, and is quite different from that to be advocated in what follows' (Russell 1905, p. 42n.* ).Russell is rewriting history but this suggests that at least in his mind the theory in Russell 1903 can do what Frege's theory does, that is, furnish mental contents that are not the things themselves.Beyond this he seeks to distinguish himself from Frege.
As soon as this epistemological concession is made, transparency ceases to be viable as an approach to meaning, because the mind acts in some sense as a medium, sorting knowledge into categories of acquaintance and of description.The analysis of a proposition can no longer be checked by pairing off words and terms, because knowledge by description can take a different logical form.What has also been lost is Russell's earlier clear-cut distinction between the logical and the psychological.
). 26 Arguably Russell has a modest or austere notion of sense construed as transparent representation, with the view he attributes to Frege being immodest or full-blooded.But this would take us too far afield.is said, and what there is.This is hard to state because there is no view from sideways on.There is no way of saying, here is what there is, here is what we think and say about it.There is no sensible way of stating a distinction, because they stand in line, each equivalent to each.This also leaves Russell without a means of accounting for the potential informativeness of statements of identity, for adding to George IV's stock of information by informing him that Scott is indeed the author of Waverley.Fregean senses won't do, because if one is to know what one is talking and thinking about they can only be transparent and hence superfluous intermediaries.The 1903 theory of denoting doesn't help because it is a purely logical theory, with no connection to our cognitive capacities ('the author of Waverley' is a denoting concept because it includes the syncategorematical term 'the').Faced with Evans's intuitive criterion of difference, in 1904, Russell's response would have been (a) it hasn't been shown how senses are supposed to do this, and (b) this looks to be lapsing into psychology.If one seeks to keep knowledge within the domain of the logical, the difficulty remains.Nevertheless the argument put forward here gives Russell an answer to the question, 'how do we think of, and speak of, what there is?', and does so without committing him to the prima facie absurd claim that in thinking a proposition one somehow has the things themselves in mind. 27