C. D. Broad, Critical Notice of A. Meinong, Über Annahmen (Leipzig, 1910). Mind, n.s., v. 22, January 1913, pp. 90-102.

VI -- CRITICAL NOTICES.

Über Annahmen.Von A. Meinong. Zweite umgearbeitete Auflage. Published by J. A. Barth. Pp. xvi + 403.

Some apology is needed for the tardiness of this notice of a work which bears the date 1910. The reviewer can only plead that the book did not fall into his hands till late in last year and that it deserves something better than a hurried reading.

The new edition of Über Annahmen is considerably larger than the first, as a good deal of controversy has raged over the subject and Meinong has taken it into consideration. In England there have been three important articles by Mr. Russell in Mind, vol. xiii., and abroad there has been war to the knife with Marty. Meinong has also made changes of arrangement, and, on certain points, changes of view.

Everyone is or ought to be acquainted with the thesis of Meinong's extraordinarily able and important work. It is that beside acts of judgment and ideas there is an intermediate kind of psychical state -- the act of supposing -- which resembles judgment in that its content can be affirmative or negative, but differs from it and resembles ideas in that it is unaccompanied by conviction. Meinong tries to show that it is necessary to assume such acts for a variety of reasons and that they throw a light on some of the most; difficult questions in the theory of knowledge. The extreme value of the book lies not merely in the evidence brought forward for the existence of suppositions, but in the discussions to which the search for suppositions gives rise on all manner of difficult points in logic and what Meinong calls 'Gegenstandstheorie'. There is further a contribution to Ethics and Aesthetics in Meinong's attempt to show the necessity of assuming something comparable to suppositions in the realms of Feeling and Volition.

Meinong does not think it necessary to prove that suppositions differ from judgments, but he thinks that he must prove that they differ from ideas. He considers that he himself has introspective evidence for this difference; but he admits that it is better to have a proof. It will be remembered that Mr. Russell saw no reason to differentiate between the two. Meinong's argument on this point is important, for he constantly appeals to it throughout the book. It runs as follows.

We can suppose negative propositions, but a negative cannot be, grasped by an idea, but only by something like a judgment. The latter point he attempts to prove, and, to do this, he has to assume certain characteristic conclusions of his theory of Objects of Higher Order. A negative is certainly a complex; and the idea of a complex, though certainly not in any sense the sum of the ideas of the elements, is yet 'produced by' these ideas and cannot occur without them. Hence if there were an idea of not-A there must be ideas of A and of something else to serve as foundations for this idea. At this point Meinong discusses the suggestion that propositions of the form A is not B can be reduced to ones of the form A differs from B. Difference is the object of a produced idea; and so, if negation could be reduced to assertion of difference, it would be plausible to hold that there are produced ideas of negatives.

Meinong has a general argument which, if valid, would be fatal to any attempt to make negatives objects of produced ideas. It is as follows. The judgments based on produced ideas of complexes and asserting the relation of their elements are a priori and necessary. If negatives can be presented by ideas it must be by produced ones, and negative judgments must be necessary. But many negative judgments are not necessary. This argument does not seem to me satisfactory. Take Meinong's examples. It is necessary that red differs from blue, but contingent that a stone let go does not rise from the earth. Hence the examples seem in his favour. But take the proposition: the Vice-Chancellor of Cambridge University in 1912 differs from the Master of Trinity in 1912. This proposition is true and is about difference which is an object of higher order; yet it is surely as contingent as the one about the stone. If this be so the fact that some negative propositions are contingent is no ground for denying that negatives may be objects of produced ideas.

The question whether negatives can be objects of produced ideas seems then to remain open. But we may glance at Meinong's special arguments against the reduction of negation to difference. Meinong holds that you cannot identify 'is not' with 'differs from,' because difference has degrees whilst A is not B is a statement incapable of degrees. Again he thinks that such an interpretation clearly breaks down over propositions that deny existence: 'Perpetual motion does not exist' cannot be the same as 'Everything that exists differs from perpetual motion'. Taking the second point first we may agree that although the two propositions are equivalent they are not identical. On the other hand if we take the form: 'A perpetual motion differs from any motion that exists,' it is not so clear that this proposition differs from what we are thinking about when we assert that perpetual motion does not exist. As the word 'is' is so ambiguous we may fairly expect 'is not' to have several different meanings. The interpretation of 'is not' by 'differs from' is most natural where 'is' asserts identity as in 'Mr. Asquith is the Prime Minister'. If it is to be valid elsewhere we must suppose that all other meanings of 'is' can be reduced to assertions of identity in some respect. The most common use of 'is' is to express inherence as in 'the pillar-box is red'. You cannot deny this by asserting that the pillar-box differs from red, for this is so whether it be red or not. On the other hand 'the pillar-box differs in colour from everything that is red' is the denial of ' the pillar-box is identical in colour with something that is red,' which is certainly not what was meant by asserting that the pillar-box is red. So I think we may agree with Meinong that not all negations can be reduced to assertions of difference. At the same time his argument that difference has degrees does not seem to me valid since 'difference in some degree,' which is what the proposed substitution has in mind, has no degrees.

In Chapter V. Meinong has an argumentum ad hominem against Mr. Russell's view that supposition may be merely ideas. His argument is that Mr. Russell admits that in judgments there is a difference of content according as the proposition judged is positive or negative. But there cannot be such a difference in ideas. This does not seem to me a strong argument even ad hominem. If two sorts of act can grasp the same object it does not follow that because in acts of one kind there is a difference of content corresponding to differences in the object there must be differences of content in acts of the other kind. Moreover I do not see why it should be certain that there are not such differences of content in ideas, in view of the notorious difficulty of discovering anything about content by direct introspection. Finally I do not think that it is nearly so certain that there is a difference of content between the supposition of P and the supposition of not-P as it is that there is such a difference in the corresponding judgments. I cannot help thinking that there are really three different attitudes towards a proposition and that Meinong confounds two of them under the name Annahme. These two I would distinguish as supposition and entertainment. It seems to me that entertainment clearly differs from supposition and is presupposed both by it and by judgment. When Meinong insists on the resemblance of Annahmen to judgment I think he has suppositions in mind; when he says that every judgment presupposes a corresponding Annahme I think he has entertainment in mind. But entertainment as distinct from supposition does not seem to me to differ from having an idea.

In the second chapter Meinong considers the characteristic function of sentences (Sätze). It must be noted that by these he means noises or marks on paper of a certain kind and neither judgments, which name he restricts to a class of mental acts, nor the objects grasped by such mental acts. The latter indeed are often called judgments or propositions, but we, following Meinong, will call them Objectives.

Of sentences it may be said 'that they are expressions and have meaning. These noises or marks allow us to infer the existence of certain psychical states (e.g., judgments). The judgments then have the sentences for their expression. These psychical states, further have objects, and these objects are the meanings of the sentences.

The example of the sentence illustrates Meinong's general theory of expression and meaning. But he introduces further refinements. You can sometimes infer from a sentence which expresses a judgment the existence of other psychical states. If a man says: 'I have toothache,' and you believe him you can infer the existence both of a judgment and a feeling in his mind. The sentence is then primarily the expression of a judgment and secondarily of a feeling.

Beside secondary expression Meinong also introduces secondary meaning. This depends on the theory of introspection which he put forward in his book Über die Erfahrungsgrundlagen unseres Wissens. Certain states of mind like judgments from their very nature have objects, or, as he says, 'present'. Others, like feelings, do not have objects, but can, on his view be made to present themselves. In this case they become their own objects, and then the word or sentence that expresses them gets (what it lacked before) a meaning. This he calls 'secondary meaning'.

All this discussion is preparatory to the question whether sentences always express judgments. Meinong holds that this is false both of principal and subordinate sentences. The principal sentence: 'Is it raining?' does not express a judgment; for, if we judged either that it was or that it was not raining, we should not ask the question. Similarly in 'I am uncertain whether Smith is trustworthy' I make no judgment about Smith's dependability. The mental act expressed in all such cases is an Annahme. Finally when a man uses a sentence which expresses a judgment and we understand him we do not as a rule either make a judgment like his or make a judgment about his judgment. We simply make an Annahme with the same objective as his judgment.

The third chapter deals with Objectives. These are the direct objects of acts of judging and of Annahmen. They are objects, of higher order. In general a judgment or Annahme needs a presentation of an object as its foundation. This object we can if we like call the indirect object of the judgment. But it is best, to say that the objective is what is judged and the indirect object what is judged about. Take the judgment that grass is green. What is judged is 'that grass is green'; and this is the objective. But the judgment is founded on ideas of grass and of green. And grass, which is what is judged about, is the indirect object. But the objective of one judgment or Annahme can become the indirect object of another. In 'it is certain that grass is green' what is judged to be certain is neither, grass nor green but 'that grass, is green'. Hence 'that grass is green' which is the immediate object or objective of the judgment 'grass is green' is the indirect object of 'it is certain that grass is green'. It is clear that when objectives become indirect objects of fresh judgments they must often be presented by Annahmen and not by judgments. This is obvious in such a case as: It is false that 2 + 2 = 5.

Objectives do not exist but subsist. And they are timeless. Meinong has no difficulty in showing that arguments against the latter view rest on a confusion between the time involved in the objective and the time at which the judgment of it happened. It is important to be clear on the relation between subsistence and truth. Apparently all subsistent objectives are true; for he calls them facts. On the other hand some subsistents are not capable of truth or falsehood; e.g., the difference between red and blue subsists, but it is neither true nor false. And it looks as if false objectives, though they do not exist or subsist, must have some third kind of being. Yet it will be a kind that has no negative. Meinong refuses to come to a definite conclusion about it and decides to call it by the non-committal name of Aussersein.

Another characteristic of objectives is that they have modal qualities. It is perfectly true that there are also differences in the correlated contents of the acts that grasp objectives and that we may reach differences of modality by reflecting on these acts. But we can and do generally learn about the modality of an objective by inspecting the objective itself. Meinong's own account of modality is complicated, and I do not feel confident that I have understood it.

He distinguishes certainty and evidence in the judgment and says that they correspond to actuality in the objective. He then tries to prove that certainty belongs to the act and evidence to the content. For, he says, that belongs to the content of a psychical state which cannot change while the object remains the same. Now certainty can change in degree whilst the objective remains the same. But evidence belongs to content; for an evident judgment cannot grasp any but an actual objective. Yet further refinements are introduced. There is evidence for certainty and evidence for probability. The former alone corresponds to actuality, the latter to possibility in the objective. And again there is another kind of evidence, viz., Eational Evidence which corresponds to necessity in the objective.

I find it difficult to see how a theory which accepts objective necessity can admit possibility as a quality of objectives. If a proposition be true its falsity is impossible. If it be false its truth is impossible. But it must be either true or false. Hence for any proposition either its truth is impossible (and its falsehood therefore necessary) or vice versa. Where then is there room for objective possibility if objective necessity be granted? There is also a difficulty about evidence and certainty. It is clear that both are meant to be psychological. Yet the phrases 'evidence for certainty,' etc., suggest that there is something purely logical about evidence. Again evidence is said only to exist in judgments that grasp actual objectives. If this be so we ought to be able by careful enough introspection to determine that a true judgment is true (though not that a false one is false). Meinong in fact seems to hold that a truly-evident judgment cannot have a non-actual objective; but this is one of those statements which are not very helpful because if false they could not be refuted. For if an apparently evident judgment came to be doubted Meinong would merely have to say that it had never really been evident. The mixture of logical and psychological elements in certainty and the evidence for it which we noted above is not necessarily a fault in Meinong, but is something typical of this very difficult subject. On the one hand certainty is purely psychological and can exist in any degree with the same objective; on the other hand there is a right degree of certainty which depends on the nature of the objective judged and on other objectives. The mystery of this state of affairs is not lessened by introducing a kind of logico-psychological hybrid in the shape of evidence for certainty or for probability between the logical subsistent objective and the psychical existent act.

In Chapter IV. are enumerated these cases of Annahmen that can be found by direct inspection. He finds them in games, art, lies, questions, and desire.

In Chapter V. Meinong considers what kinds of acts can present objectives. Of course judgments and Annahmen can do so; but not, he thinks, ideas. This question we have discussed earlier, but there remain a few points to notice. He insists on the extreme difference between the objects that are admitted to be objects of ideas and objectives; e.g., between a mountain and the existence of a mountain. But does not this difference mainly lie in the fact that a mountain can be the object of a perception whilst the existence of a mountain cannot? But not all objects of ideas are perceptible. The British Constitution can presumably be the object of an idea, even if, as we shall see later Meinong holds, an Annahme be also needed to grasp it. Yet the British Constitution is not so different frem the existence of a mountain as both are from a mountain.

Meinong next tries to prove that objectives cannot be indirectly grasped by ideas. By this he means that if you try to grasp objectives through descriptions such as 'the objective which is grasped by the judgment J,' you will still need something more than ideas. For you will need a direct acquaintance with the judgment J; and J, being a psychical state, cannot be the object of an idea. This last opinion depends on Meinong's theory of introspection already mentioned.

According to Meinong we do not desire objects but objectives. When we say that we desire X we really mean that we desire X's existence. And such objectives must be grasped by Annahmen and not by judgments. If we judged that X existed we should not desire it, though we might of course desire its continuance. But could we not be said to desire at t1 the existence of X at t2 although we judge at t1 that X will exist at t2? If not, it will follow that nothing which we believe will exist if and only if we desire it ever will exist. For if we hold this belief and desire it we shall believe that it will exist, and then we shall cease to desire it and it will never exist if our belief as to the conditions of its existence be true. There is too a further difficulty about the doctrine that we desire objectives which Meinong does not seem to notice. When we say that we desire X it is reasonable to hold that what we really desire is the existence of X, and it is true that the latter is an objective. Yet X never may exist. In that case 'the existence of X is not actual. If false objectives do not subsist in some sense it will follow that we have literally desired nothing unless our desire is one that will be fulfilled. If on the other hand they do subsist we seem forced to say that it is not the existence of X that would satisfy us but the actuality of X's existence; hence it is the actuality of X's, existence that we really desire. But the actuality of X's existence may itself be false (and will be so if X's existence be not actual); hence we seem to have started on an infinite regress in trying to state what we really desire.

The next chapter which deals with Operations on Objectives is very important, for it is largely concerned with the nature of inference. When we infer q from p 'persuasion' is in some sense conveyed from Jp to Jq. This however cannot simply mean that the judgment Jp is a part of the cause of Jr. For this may be the case when we do not say that we have inferred q from p. Moreover we directly perceive that p is the ground for q, whilst we can only find out by experiment what causes anything. Meinong compares the conveyance of persuasion from one judgment to another to the production of an evident judgment of comparison by the mere presentations of the terms compared. In inferring 'A is C' from 'A is B' and 'B is C' the judgments of the last two objectives stand in the same relation to the judgment of the first as do the presentations of X and of Y to the evident judgment of 'X differs from Y'. This peculiar relation is expressed by saying that we judge A is C 'in view of' (im Hinblick auf) our judgments of the other two objectives. The experience of 'judging in view of' is ultimate. He compares this relation to that between desires when we desire the means in view of an existing desire for an end.

But sometimes when Jp is evident we believe Jq to be evident when it is really false and some fallacy has been committed in inferring q from p. And again when we do not think Jp or Jq evident a formally correct inference of q from p seems to give some evidence to q. We say at least that it is evident that q really follows from p. As evidence for Meinong implies truth he cannot count this as real evidence, but calls it quasi-evidence. Here there is considerable departure from the First Edition where quasi-evidence was called relative evidence, and true evidence was treated as a special case of it. Meinong's present position is that in formally correct arguments when the truth of the premises is not asserted both premises and conclusion are angenommen, but that Annahmen can have mediate evidence as well as judgments. There are great difficulties about this view and Meinong discusses them on page 350.

He has to suppose that Annahmen can have mediate but not immediate evidence. And this is in great contrast to judgments which only get their mediate evidence from being judged in view of others that are immediately evident. Meinong's reply is that of the two elements in mediate evidence (viz., evidence of premises and judgment in view of them) only the analogue of the latter may be needed for Annahmen. In fact an Annahme can become evident in view of another that is not itself evident. This seems to me a difficult position to maintain. But if we distinguish supposing from entertaining I think we may fairly hold that whilst entertainment has no evidence, suppositions can have both immediate and mediate evidence. But there seems a general difficulty about Meinong's theory of evidence. If we remember that evidence belongs to content we see that it must be uniquely correlated with some quality in the objective: -- degree of possibility presumably. The latter cannot alter. Yet if we judge the same objective in view of premises of various degrees of immediate evidence the mediate evidence of the judgment of the conclusion will vary. And so the correlation between the evidence of the judgment and the degree of possibility in the objective disappears.

Meinong rejects the view that a piece of reasoning like a syllogism which would be an inference if its premises were asserted is nothing but a hypothetical proposition with a complex antecedent when they are merely angenommen. His ground for this seems to be that we say 'suppose M is P' and 'suppose S is M' (expressions indicative of Annahmen) but add 'then S is P'. If this distinction is to be maintained it ought to be applied to hypotheticals with simple antecedents too. The true hypothetical ought to be 'if A were B C would be D,' and the non-inferential form of the syllogism 'if M were P and S were M then S would be P'. But I doubt if these verbal distinctions express any real logical difference. We tend to use the latter form when we believe that S is not P, and of course it is true that this is a case where S is P can only be angenommen, and where inference is useless since we know the relation of S to P apart from the syllogism. It is the latter consideration that really distinguishes the two forms. In both cases where the premises are merely angenommen there is no inference and the conclusion is angenommen too. But when it is expressed in the form S is P we mean that we are ready and willing to pass from Annahme to judgment if we get the chance to infer; whilst, in the conclusion S would be P, we mean that we have already made up our minds that S is not P and do not want or expect to infer anything.

We now come to Meinong's treatment of the hypothetical judgment itself. He says that if it be an ordinary judgment it is strange ihat it has no contradictory. But what he calls strange seems not to be true. The contradictory of If p then q is Though p yet not q. Meinong's own view is that in a hypothetical judgment what is really asserted is the consequent as modified by the antecedent. 'If a triangle be isosceles the angles at its base are equal' becomes 'The angles at the base of an isosceles triangle are equal'. This substitution is an old friend, and I do not see that Meinong has cured it of any of its weaknesses. Of course it only applies at all well to conditionals as distinct from true hypothetical, i.e., its propositions of the form φx)xψx as distinct from p)q. If we apply the substitution to 'If it rains I shall get wet' we obtain 'I who am in the rain will get wet' -- a proposition which either fails to express my meaning or must reintroduce a hypothetical if I am in the dry. Meinong is more troubled about existential hypotheticals. If you reduce 'if Gods exist divine works exist' to 'the works of existing Gods exist' there is the same difficulty as before in case no Gods do exist. Meinong's only solution is to point to other existential propositions about non-existents : e.g., ' an existent round square exists'. This proposition, he says, is true, although round squares do not exist. A contradiction and a round square seem a flimsy basis for a theory of the hypothetical judgment.

The theory of the modified consequent does not however exhaust his account of hypotheticals. He thinks that the modified consequent is asserted in view of the antecedent which supplies conviction, or, in some cases, evidence. There seems to be a difficulty in reconciling this with some that has gone before. Hypotheticals are not always judged, they are often merely angenommen. By analogy this must mean that the modified consequent is angenommen in view of the antecedent, and gets some evidence from it. But we can suppose the proposition 'if a triangle be isosceles the jangles at its base are unequal'. Can we possibly hold, as we must do on Melnong's theory, that the supposition that the angles at the base of an isosceles triangle are unequal gains evidence when supposed in view of the equality of its sides?

There is more to be added, however. If the hypothetical judgment be only an operation ending in a categorical judgment why call it a special kind of judgment ? Is there no special experience which the verbal form expresses and whose object is its meaning? Meinong thinks that there is. The various experiences themselves which grasp the objectives and end in the Judgment can be used too to grasp an objective of higher order. This is the complex of the two objectives related by the 'if-relation'. The hypothetical judgment is not itself the recognition of an if-relation, but the meaning of a hypothetical proposition is a complex related by that relation and capable of being grasped.

In the IXth Chapter Meinong considers the general question of the presentation of objects. He thinks that Annahmen are largely concerned in this. He holds that in the First Edition he was obsessed by a prejudice in favour of the existent. He therefore reproduces what he said there with the warning superscription of Existential View and adds his corrections under the heading of Non-existential View. The existential view is that the mediate objects of true affirmative existential judgments exist. A false affirmative existential judgment has no mediate object; but we may call its mediate object that which it would have if it were true. A similar expedient is adopted for true negative existential judgments. But in this case how can we strictly say that all judgments have mediate objects? His first suggestion is that the objects are presented by ideas, and that ideas have nothing to do with truth or falsehood. Still the ideas on which a true negative or false affirmative existential judgment is based will have non-existent objects. He has to overcome this difficulty by the notion of potential objectivity. This must be a psychical disposition. But then a disposition, though it is something, is not something that is presented, whilst every idea does seem to present an object. His final solution on the existential view is as follows. When we make a positive existential judgment we find on introspection the experience of grasping a mediate object, whether the judgment be true or false. Why not suppose then that this experience is always due to the existence of something like a judgment? When our judgment is a false affirmative or a true negative the experience of grasping an object is due to the existence of a positive Annahme. Further we must suppose that we only experience an idea as presenting an object when it is followed by an Annahme that the object exists. If this be so it will explain how contradictions like round squares can apparently become mediate objects; for Annahmen are indifferent to contradictions. And finally the very fundamental character which positive Annahmen now assume is compared with the essential positivity of the suggested Aussersein.

I shall not criticise the above theory, but will pass at once to the Non-existential View which Meinong now holds. The existential view held that ideas of existents actually grasp existent objects, and tried to explain the experience of grasping an object in cases where what seems to be grasped does not really exist. The present view maintains (a) that every judgment has a mediate object whether that object exist or not. This amounts to a reiteration of the commonly accepted view which we express by saying that it is necessary to 'know what we are talking about'. But (b) it holds that no idea alone ever grasps an object even when the object exists. Having an idea is a passive state, whilst grasping an object is an action; hence the former is not enough for the latter. And it is clear that the kind of ideas called sensations very often do not grasp objects although they can be used for that purpose. Finally Meinong uses an argument based on his theory of introspection. The content of an act and its object are uniquely correlated. But, if Meinong's theory of introspection be true, the content of an idea can be used both to grasp its so-called object and to grasp the idea itself. Hence in at least one case something must be added to the content of the idea if the unique correlation of content and object is to be retained.

This addition Meinong calls 'intending' (Meinen). I intend X by means of the positive Annahme that X exists or that X subsists. If an existential judgment be affirmative and false or negative and true still the objective of this Annahme subsists and the object X is grasped in it. There seem to be two difficulties in this theory. Firstly it does not help us over non-subsistent objects. Suppose that the objective that X subsists be false then, though it may be true that the objective has some kind of being and can be angenommen, this does not bring us any nearer to intending X, for there seems to be no X to intend. In fact an objective asserting that a non-subsistent object subsists must be itself false. If false objectives have some kind of being so may non-subsistent objects and the by-way through the objective is unnecessary. If false objectives have no kind of being then the expedient is useless, for how can they be grasped? The other difficulty is more general. Is it not just as necessary to grasp an object in order to make an Annahme about it as in order to make a judgment about it? If so the theory involves a vicious infinite regress of positive Annahmen.

In Chapter VIII. Meinong considers the difference between intuitive and non-intuitive ideas. Whenever you have an intuitive idea you have a complex object. Now you can have a non-intuitive idea of the same object. Hence the difference must lie in the fact that the contents of the ideas of the elements of the complex are differently related according as your idea of the complex is intuitive or non-intuitive. When the idea is intuitive Meinong calls the contents of the ideas of the elements unified (zusammengesetzt). When it is non-intuitive he calls them united (zusammengestellt). Now a non-intuitive idea whose elements are the ideas of X and of Y can grasp either the object X that is Y or the object X that is not Y. But an intuitive idea can only grasp the former. Further no idea alone can grasp the latter; a negative judgment or Annahme is needed. So that the final distinction is that an intuitive idea of a complex is one that can only give rise to an Annahme or judgment asserting one element to inhere in the other, whilst a non-intuitive idea can give rise to either a positive or negative judgment or Annahme.

Meinong chooses to treat separately as a more complicated case the question of the presentation of two terms in relation, e.g.,.red differing from blue. Whether there be any difference between this and the earlier cases depends on whether inherence be an ordinary relation. He argues here that, although ideas of red and of blue and of difference are necessary to present red differing from blue, they are not sufficient. He uses an argument familiar to readers of 'The Principles of Mathematics' about the distinction between a relation as such and a relating relation. But he tries further to prove the general proposition that if a number of contents separately be not adequate to a given object no combination of them can be so. By a content being adequate he means that it gives rise to and justifies a statement about the object. Unfortunately he does not tell us what the statement must be; but we may fairly suppose that in the example it is that red and blue differ. Now the relation between content and object, he says, is an ideal relation, and those between contents real ones. An ideal relation is one which alone can relate terms that subsist but do not exist, though it can relate existents too. It is a property of such a relation that if it relates existents it only ceases to hold through changes in its terms and not through changes in their real relations to other existents. By definition the relation between content and object is ideal when the object only subsists. (Meinong seems to think that it also follows when the object is an existent; but this is of course only a plausible assumption.) Hence he argues that if single contents be not adequate to a given object no alterations of their real relations will make them so.

This argument does not seem to me to be cogent. It only proves that the separate contents will not become adequate through changes in their real relations, and not that a complex of these contents related by certain real relations may not be adequate to an object to which none of them separately was adequate.

At the end of this chapter Meinong distinguishes two kinds of intending. You may intend an object not merely by supposing or entertaining the objective that it exists or subsists, but by doing the like with objectives that assert qualities of it. He calls the former Seinsmeinen and the latter Soseinsmeinen. We may call them direct and indirect intending respectively. So far as I can see indirect intending corresponds closely to what Mr. Russell calls knowledge of description. But what exactly is the objective angenommen when we indirectly intend an object? It is clear that it must be a proposition. On page 273 Meinong calls 'the mountain is golden' the objective by supposing which we indirectly intend 'the golden mountain'. By analogy I take it that the objective angenommen in indirectly intending 'the discoverer of Radium' would be 'he discovered Radium'. But the phrases 'he' and 'the mountain' are obviously incomplete. We ask immediately: Who? and What mountain? And these are just the questions that ought not to arise if by supposing these objectives we have indirectly intended the objects. Surely what must be angenommen is not 'x is golden and a mountain' or 'x discovered Radium,' but 'there is an x such that x discovered Radium' and 'there is an x such that x is golden and a mountain'. But then we have got back to Seinsmeinen.

I shall say very little about the IXth Chapter, because it is of less general interest than the others, consisting as it does largely of a polemic against Von Ehrenfels' views of desire and value. it is only necessary to note that Meinong holds that there are psychical states which stand in the same intermediate position between ideas and desires or ideas and feelings as do Annahmen between ideas and judgments. He holds further that there is a general law about the causation of desires, which runs as follows. In desire we present an object, we then suppose the objective that it exists. This Annahme causes a quasi-feeling, and, if the latter be pleasurable it causes us to desire the existence of the object.

The last chapter consists of a summary of the results of the 'work. The book as a whole can safely be described as a model of acute and profound investigation into the hardest and most fundamental questions of philosophy.

C. D. Broad.