Four Theses in Frege (penultimate draft)

PIERANNA GARAVASO*

FOUR THESES IN FREGE1

1. Introduction

There are two interesting debates concerningfour theses attributed to Frege. In both cases,the debate arises from the apparent inconsistencybetween two theses for each of which many scholarshave claimed there is strong textual evidence inFrege’s writings. Two of these theses areconnected with one another and might even be takenas distinct versions of the so calledCompositionality Principle, i.e., the thesis thatthe composition of sentences expressing thoughtsparallels the composition of the thoughtsexpressed by such sentences. These alleged Fregeaninconsistencies have generated lively but so farindependent debates as these puzzles have not beendiscussed in connection with one another. In thispaper, I briefly outline the puzzles and attemptto begin a simultaneous discussion of thesetheses. My goal is to show that the epistemicsolution I provide for one puzzle allows aplausible integration of all these theses. Isubmit that one promising way to connect these

** University of Minnesota Morris. [email protected] I would like to thank Lory Lemke for feedback and discussions that have helped improve the paper.

1

debates is by stressing how for Frege languageplays an indispensable epistemic role in thegrasping of thoughts.

2. Four Theses and Two Puzzles

The first puzzle arises from an alleged conflictbetween the following two theses:

(A) Distinct sentences can express the same thought (B) There is a correspondence between the parts of a

thought and the parts of the sentence expressing thisthought

The second puzzle arises from the following twotheses:

(C) The Context Principle: «[N]ever to ask for themeaning of a word in isolation, but only in the contextof a proposition» (Frege, 1884; Eng. tr., p. X).

(D) The Compositionality Principle: «the meaning of acomplex expression is fully determined by its structureand the meanings of its constituents» (Szabó 2012).

The discussion within the debates concerningthese puzzles assumes that Frege was committed tosome version of these theses during most of thedevelopment of his thinking and not only during alimited period. For the sake of discussion andbecause, in my opinion, the textual evidenceprovided by the scholars engaged in these debatesis impressive, I assume that Frege was indeedcommitted to all these theses and that they playedan important role throughout the whole developmentof his thought.

2

3. Thesis (A): Distinct sentences can express the same thought

Frege mentions several examples of pairs ofdistinct sentences expressing the same thought andthese pairs represent dissimilar cases2. Forexample, the active and passive forms of asentence such as «The Greek defeated the Persiansin the Battle of Platea» and «The Persians weredefeated by the Greeks in the Battle of Platea»have the same sense, i.e., express the samethought (Frege, 1879; Eng. tr., p. 112). Fregediscusses a different case in the followingpassage:

In the sentence ‘there is at least one square root of4’, we are saying something, not about (say) thedefinite number 2, nor about -2, but about a concept,square root of 4; viz, that it is not empty. But if Iexpress the same thought thus: ‘The concept square root of4 is realized’, then the first six words form theproper name of an object, and it is about this objectthat something is being said (1892b; Eng. tr., p. 188).

In the case with passive and active sentences,the grammatical structure of the two sentences ismore similar than in the second case, where thesentences cannot be parsed into the same partssince one sentence is of the subject-predicate2 Silver Bronzo lists eight kinds of relevant and non-relevantcases in the so far most comprehensive listing of examples ofThesis (A), which he calls Frege’s Multiple Analyses Thesis. See Bronzo (manuscript). Another extensive listing is in Kemmerling (2011, pp. 175-176).

3

form and the other is an existential statement.Accordingly, for Frege, even sentences with quitedifferent structures may express the same thoughtcontent3.

4. Thesis (B): There is a correspondence between the parts of athought and the parts of the sentence expressing this thought

Frege repeatedly stated that the parts of athought somehow parallel the parts of the sentenceexpressing this thought. Here are two laterstatements of (B):

The world of thoughts has a model in the world ofsentences, expressions, words, signs. To the structureof the thought there corresponds the compounding ofwords into a sentence (1918-1919; Eng. tr., p. 38).

We can regard a sentence as a mapping of a thought;corresponding to the whole-part relation of a thoughtand its parts we have, by and large, the same relationfor the sentence and its parts (1919; Eng. tr., p.255)4.

Wolfgang Künne dubs (B) ‘a principle of Thought-Sentence Correspondence’ stressing the parallel

3 For other relevant examples, see Frege (1899-1906?; Eng. tr., p. 167; 1906a; Eng. tr., p. 187; and 1924-1925; Eng. tr., p. 269).4 See also «As a sentence is generally a complex sign, so the thought expressed by it is complex too: in fact it is put together in such a way that parts of the thought correspond to parts of the sentence». (1914; Eng. tr., pp. 207-208, 225,243); and again in Frege (1923; Eng. tr., p. 55).

4

between the structure of thoughts or senses andthe structure of sentences (Künne 2007, p. 95).It would be interesting and exegetically veryuseful to identify all the versions of (B) inFrege’s writings and in the secondary literatureon Frege; it would also be plausible to argue thatnot all of them are synonymous.

Although he often used the part/whole relationin describing the link between thoughts and theirparts, Frege was not satisfied with its use; hemade clear that this relationship should be takenmetaphorically: «To be sure, we really talkfiguratively when we transfer the relation ofwhole and part to thoughts; yet the analogy is soready to hand and so generally appropriate that weare hardly ever bothered by the hitches whichoccur from time to time» (1923, p. 55).Furthermore, in these passages and in the textsfrom which these passages have been extracted,there is no reference to senses and meaning ofsentences; hence, these formulations of the ideaof the compositionality of thoughts do not haveany explicit connection with semantic issues. Bothpassages were written well after Frege’sintroduction of the distinction between sense andreference in his Ṻber Sinn und Bedeutung in 1892;accordingly Frege thought of the parallels betweenlanguage and thoughts not only in terms of thelinking of senses to words as (D) explicitlyrequires.

The wording of the statements of (B) and theappeal to the part/whole relationship have causedsome of the discussion of the first puzzle tofocus on the issue whether or not thoughts possess

5

unique and essential internal structures5. Thisquestion is related to the question of whetherthere are links between the complete sense of asentence and the senses of the words composing it,but it is not the same question. Even if an answerto the first question will have inevitablerepercussions for the potential answers to thesecond question, it is useful to keep clear thedifference between the original questions thathave given rise to the two distinct debates.

5. The First Puzzle

The puzzle between (A) and (B) can be stated asfollows. Assuming that we have a satisfactory wayof determining how the parts of sentences andthoughts are understood, how can the parts of oneand the same thought correspond to the parts ofdistinct sentences? Consider for example the pairof sentences mentioned by Frege in the passagequoted in section 3, i.e., «There is at least onesquare root of 4» and «The concept square root of 4 isrealized». Even if we agree on looking at the

5 The discussion on whether or not Frege was committed to the view that thoughts have unique essential structure originatedmostly from Dummett’s distinction between decomposition and analysis and his claim that for Frege, analysis is unique and «displays … the ‘essential structure’ of the sentence, and may be said to uncover the internal structure of the thought it expresses» (1981, 272). Andreas Kemmerling lists several scholars supporting this view: «David Bell (1987), Dalia Drai, Michael Dummett, Paul Horwich, Jeffrey King, Christopher Peacocke, Ian Rumfitt, Stephen Schiffer, Robert Stalnaker, and Pavel Tichy». (2011, p. 166).

6

grammatical structure of the sentences todetermine their components, it is prima facieclear that any analysis of these two sentenceswill lead to different components: how can “theconcept square root of 4”, which is a component of thesecond sentence correspond to any part of thethought expressed by the first sentence, whenthere is no component in the first sentence thatcorresponds to “the concept square root of 4” and yetthe parts of this first sentence are supposed tocorrespond to the parts of the thought expressedby both sentences?

Silver Bronzo (manuscript) usefully parts themany solutions offered for this puzzle into twocamps, namely, the Dummettian and the Anti-Dummettiancamps, which can be characterized by the rejectionof one of the two theses from which the puzzleensues. The Dummettian camp claims that Frege iscommitted to some version of (B), but not to (A);the Anti-Dummettian camp on the other side claimsthat Frege endorsed (A) but not any version of(B). Bronzo’s goal, which he shares with otherscholars, who cannot be listed with either of theabove camps, is to maintain that Frege held boththeses and that there is a Fregean solution of thealleged conflict6.

6. Thesis (C): The Context Principle

6 Among these, see H. T. Hodes (1982), Gregory Currie (1985), Pieranna Garavaso (1991), José M. Bermudez (2001), Carlo Penco (2003), Mark Textor (2009), and Andreas Kemmerling (2011).

7

The fact that Frege placed this principle in aprominent position in his Introduction to theGrundlagen tells us that it was a principleimportant for him. It helps to read this principlein the context of the other two principles thatFrege mentions in the Introduction:

a. Always to separate sharply the psychological fromthe logical, the subjective from the objective;

b. Never to ask for the meaning of a word inisolation, but only in the context of a proposition;

c. Never to lose sight of the distinction betweenconcept and object.

According to most Fregean scholars, the ContextPrinciple appears for the first time in theIntroduction to the Grundlagen7; there seems to bealso a prevalent consensus on the fact that thecontext principle does not appear in any otherwriting by Frege, although there are fewdiscordant voices. David Bell (1981, p. 211)finds a first statement of the principle in thefollowing passage from the Begriffsschrift:

What is asserted of the number 20 cannot be assertedin the same sense of [the concept] “every positiveinteger”; … The expression ‘every positive integer’ byitself, unlike [the expression] ‘the number 20’, givesno independent {selbständige} idea; it acquires sense onlyin the context of a sentence. (1879; Eng. tr., p. 128)

7 Although most attention has been given to this first occurrence of the context principle, Ignacio Angelelli pointsout that it occurs at least four times in the Grundlagen with three statements occurring in §§ 60, 62 and 106 (1967, p. 73).

8

Even if Bell is correct, since Begriffsschrift waswritten in 1879, i.e., before the Grundlagen,Bell’s statement does not contradict the wide-spread view that Frege supported the ContextPrinciple only in his earlier works8. In contrast,Richard Heck and Robert May suggest the presenceof the Context Principle in the Grundgesetze whenthey state:

There are significant differences between [Frege’s]statements of the context principle in Die Grundlagen (GL,p. x, §62) and in Grundgesetze (Gg, v. I, §29),differences that seem to reflect a change from a viewon which sentences have a privileged position to one onwhich they are but a kind of proper name (2011, p.159).

Theo Janssen also claims that «the contextprinciple was not repeated in [Frege’s] publishedwritings, but he did not give it up… It appears inseveral occasions in correspondence and inunpublished writings (also in one intended forpublication)» (2001, p. 133). Finally, RainerStuhlmann-Laeisz lists some additional putative

8 Since in the Begriffsschrift Frege had not yet developed the anti-psychologism of his later works, Bell’s thesis may provide further support for an interpretation of the Context Principle not so strongly anti-psychologistic as it is often interpreted. On the other side, it is Frege himself who tightly connects the Context Principle with the rejection of psychologism: «[I]f the second principle is not observed, oneis almost forced to take as the meanings of words mental pictures or acts of the individual mind and so to offend against the first principle as well» (1884, p. X)

9

statements of the context principle from bothearly and later writings by Frege (2001, p. 254)9.

Besides the issue of whether or not Frege heldsome version of the Context Principle also in thelater development of his thought, another featureof this principle is worth mentioning. It ispointed out by more than one scholar that thereare different versions of this principle.Stuhlmann-Laeisz distinguishes between fourvarieties of context principles: ontic, epistemic,sufficiency, and necessity (2001, p. 252); Marco AntonioRuffino identifies the epistemological version ofthe Context Principle with what has often beencalled “the priority thesis”, asserting thepriority of judgments over concepts (1991, p.185). I do not discuss the difference betweenvarious formulations of this principle; my attemptto connect it to the other theses relies on what Ibelieve to be the most persuasive reading of thisprinciple, namely the interpretation that takes itto be one of Frege’s most explicit epistemologicalrecommendations.

7. Thesis (D): The Compositionality Principle

This thesis that many scholars call “Frege’sPrinciple”10 focuses on the semantic contribution

9 The cited passages are from Frege (1880/1881, 1903, 1906b, 1919). Stuhlmann-Laeisz acknowledges “it has been also said that these further passages give little evidence [of the context principle], if any at all,” Ibidem,10 This denomination is rather confusing since many scholars refer to both the Compositionality Principle and the Context

10

that words as the constituents of sentencesprovide to the whole sense of a sentence andaccordingly to the expression of the thoughtconveyed by a sentence. Regarding this thesis,deep disagreement exists in the secondaryliterature concerning (1) whether or not Fregeever stated it, and (2) how the principle shouldbe interpreted.

Francis Jeffrey Pelletier provides a usefulsurvey of critical discussions of (D) and of itsattribution to Frege. He points out how acompositionality thesis may be understood asconcerning Bedeutung or Sinn or some other semanticnotion preceding Frege’s 1892 clarification of thedistinction between Bedeutung and Sinn. Pelletierargues for ascribing to Frege a referentialversion of the compositionality principleaccording to which “the Bedeutung of the whole [is]a function of the Bedeutungen of the parts” (2001,p. 104), but he ultimately argues against theattribution to Frege of a compositionalityprinciple of Sinn. The following passages have beencited to argue that Frege held a form of semanticcompositionality concerning sense:

If it were a question only of the sense of thesentence [i.e. “Odysseus was set ashore at Ithaca whilefast asleep”], the thought, it would be unnecessary tobother with the reference of a part of the sentence,only the sense, not the reference, of the part isrelevant to the sense of the whole sentence. (1892a,62-63)

Principle as Frege’s Principle; for textual evidence of this phenomenon, see Pelletier (2001, pp. 87-92).

11

The names, whether simple or themselves composite, ofwhich the name of a truth-value consists, contribute tothe expression of a thought, and this contribution ofthe individual [component] is its sense. If a name ispart of the name of a truth value, then the sense ofthe former name is part of the thought expressed by thelatter name. (1893, p. 90)

In the first quote, Frege states explicitly thatthe sense of the part of a sentence is relevant tothe sense of the whole sentence; in the second, hepoints out how the contribution that a nameprovides to the expression of a thought, i.e., toa sentence, is its sense and that this sense ispart of the thought which is expressed by the nameof the thought or the sentence. Frege proposes aclear parallel here: as the name is part of asentence, the sense of the name is part of thethought, which is the sense of the whole sentence,and the sentence in turn can be seen as the nameof the thought.

If I understand Pelletier’s denial of theascription of the compositionality principle ofSinn to Frege correctly, he does so because hebelieves that merely pointing out a correspondencebetween parts of sentences and parts of thoughtsor senses, i.e., (B), is insufficient to give ussemantic compositionality, i.e., (D). What isneeded is what Pelletier calls “the functionalityof senses”, or the claim that “the meaning of asyntactically complex expression is a function(only) of the meanings of its syntactic parts andmanner in which they are syntactically combined”

12

(2001, p. 103). Pelletier is right on this point:to ascribe to Frege a version of the semanticcompositionality principle relevant to senses, weneed to argue that Frege claimed a strongerconnection between (B) and (D). Pelletier’sposition may seem appealing because by rejectingthe thesis that Frege supported (D) we can alsoreject ascribing to Frege the alleged secondinconsistency. Despite this appealing feature,Pelletier’s position is ungrounded; there istextual evidence that Frege connected (B) and (D)more closely than the citations we have looked atso far seem to suggest.

We have noticed before that (B) and (D) primafacie focus on different notions. (B) states thatthere is a correspondence between the part ofsentences that express a thought and the parts ofthe thought expressed by the sentence; thus (B) isa thesis about the structure of thoughts andsentences and it does not mention meaning orsenses. On the other side, (D) states that thesense of a sentence is “fully determined” by thesenses of its components, or as Pelletier wouldsay, is “a function” of its components; it is thusa semantic thesis. Now, it is true that Frege didnot explicitly state that the senses of sentencesare a “function” of the senses of the componentsbut he came very close to it as he stated that,

As the thought is the sense of the whole sentence, soa part of the thought is the sense of part of thesentence (1906a; Eng. tr., p. 192).

13

And as we take a thought to be the sense of asentence, so we may call a part of a thought the senseof that part of the sentence which corresponds to it(1914; Eng. tr., p. 225)11.

These statements highlight the semantic focus of(D): the elements of the correspondence are on oneside linguistic entities such as words andsentences and on the other are thoughts and theirparts, i.e., the sense of a complete sentence andthe senses of the words composing the relevantsentence. At least with regard to thesestatements of (D), it seems clear that Frege isnot merely pointing out the structural complexityof thoughts and sentences as it seems to be thecase in the formulation of (B) and in the passagesquoted in support of Frege’s commitment to (B); inthese later passages, Frege has introduced thenotion of sense and is proposing a correspondencebetween the senses of sentences and words withthoughts and their parts; here, the correspondenceis linked to the compositional nature of thoughtsand the suggestion is made that it is because ofthe correspondence between the senses of the wordsin the sentence and the senses that compose thethought that the sense of the whole sentence isthe whole thought12. In conclusion, Pelletier iscorrect in stressing the need for a strongstatement of (D) in order to ascribe to Frege an

11 For passages cited to provide textual evidence of Frege’s commitment to (D); see Frege (1919; Eng. tr., p. 255; 1923; Eng. tr., p. 55; 1893-1903; Eng. tr., p. 90). 12 I thank Lory Lemke for helping me to read these passages asestablishing a connection between (B) and (D).

14

actual commitment to the compositionality ofsenses; however, there is textual evidence ofFrege’s support of a robust notion ofcompositionality.

8. The Second Puzzle

The second puzzle is often characterized asarising from the question of the priority ofmeaning: if the meaning of a complex expressiondepends on the meaning of its components, how canthe meaning of the components depend on themeaning of the complex expression they compose?Leila Haaparanta spells out the alleged conflictvery clearly:

The compositionality principle says that the sensesof the ingredients of a sentence S are more basic thanthe sense of S, for the sense of S is compounded out ofthem. Now, if Frege holds the view that in order tounderstand the sentence, we must understand the sensesof the words it contains, he cannot demand that inorder to understand the senses of words, we must knowthe sentences in which the words occur (1985, p. 90) 13.

It is especially helpful to state the conflictin epistemological terms: if I can learn themeaning of a word only from the context of thesentences in which it is used, as (C) states, itseems that my understanding the meaning of thesesentences should not depend on my understanding

13 For other statements of this conflict, see Dummett (1975), Fodor and Lepore (1992) and Heck and May (2011).

15

the meaning of the individual words occurring inthem as (D) states, for, if this is the case, thenthe proposed explanation of my understanding themeaning of a word would seem to be circular. Toclarify, let us characterize the Context Principleas asserting semantic holism, for a word does nothave a meaning in isolation from contexts of use,and the Compositionality Principle as assertingsemantic atomism, for the meaning of a sentencederives from the meaning of the words that composethe sentence. Each principle proposes anantithetical approach to the explanation of ourapprehension of meaning.

The solutions proposed to this puzzles have beenquite varied: there have been those who haveclaimed that the Context Principle is central toFrege’s philosophy; others who have claimed thatthe Compositionality Principle is truly “Frege’sPrinciple”. Janssen points out that in manyproposed solutions, starting with Dummett’sproposal, one of the two theses is reinterpretedor not taken literally (2001, p. 116-118).

9. The significance of (C) for the compatibility of (A), (B), and (D)

Let us review what statements we are assumingFrege held:

(A) Distinct sentences can express the same thought (B) There is a correspondence between the parts of a

thought and the parts of the sentence expressing thisthought

16

(C) The Context Principle: «[N]ever to ask for themeaning of a word in isolation, but only in the contextof a proposition»

(D) The Compositionality Principle: «the meaning of acomplex expression is fully determined by its structureand the meanings of its constituents»

By reading these theses together, can we glean acommon concern? Two threads linking these thesesare (i) Frege’s concern with the relationshipbetween language and thoughts and (ii) thechallenge of explaining human understanding ofthoughts. How can we grasp thoughts when languagemisleads us so easily, for example when we do notrecognize that distinct sentences may, despitetheir apparent differences, express the samethought? Moreover, if the composition ofsentences is so important in allowing us to graspa thought, how can we rely on sentences as thesource for our understanding the meaning of words?

Many scholars, first among them Dummett, statethat the general idea expressed by thesis (D) isthe claim that the semantic content of the partsthat compose a thought somehow “precede” thesemantic content of the thought, or «The sense of acomplex is compounded out of the senses of its constituents».Right after this statement, Dummett adds «ForFrege, we understand the sense of a complexexpression by understanding the senses of itsconstituents» (1973, p. 152). While I do notdiscuss the so called “priority thesis”, I like toemphasize the epistemological focus in the secondsentence of Dummett’s quote: many Frege’sstatements of thesis (D) stress the important role

17

that the parallel structure of thoughts andsentences plays for our apprehension of thoughtsand of the meaning of sentences. In myinterpretation and discussion of the first puzzle(1991), the structure of thoughts is crucial inFrege’s account of how we grasp thoughts; weunderstand thoughts because they have a structure,but thoughts acquire a structure once they areexpressed in language. This is especially clearin the case of new sentences that we have neverheard before; Frege states,

The possibility of our understanding propositionswhich we have never heard before rests evidently onthis, that we construct the sense of a proposition outof parts that correspond to the words (1976; Eng. tr.,p. 79).

It is remarkable what language can achieve. With afew sounds and combinations of sounds it is capable ofexpressing a huge number of thoughts, and, inparticular, thoughts which have not hitherto beengrasped or expressed by any man. How can it achieve somuch? By virtue of the fact that thoughts have partsout of which they are built up. And these parts, thesebuilding blocks, correspond to groups of sounds, out ofwhich the sentence expressing the thought is built up,so that the construction of the sentence our of partsof a sentence corresponds to the construction of athought out of parts of a thought (1914; Eng. tr., p.225).

And again,

It is astonishing what language can do. With a fewsyllables it can express an incalculable number of

18

thoughts, so that even if a thought has been grasped byan inhabitant of the Earth for the very first time, aform of words can be found in which it will beunderstood by someone else to whom it is entirely new.This would not be possible, if we could not distinguishparts in the thought corresponding to the parts of asentence, so that the structure of the sentence canserve as a picture of the structure of the thought(1923; Eng. tr., p. 55).

Passages such as the last two have led somescholars to claim that Frege believed thatthoughts have internal essential structures thatget to be reflected by the structure of thesentences expressing them . I reject thisinterpretation and submit that the apparentconflict between theses (A) and (B) disappears ifwe take Frege’s talk of parts of thoughts asfigurative as he explicitly states it is (1923, p.55, quoted before) and interpret his statementsof the compositionality principle as regarding theepistemological function of language in fosteringhuman grasping of thoughts, as in the above quotedsentence by Dummett: «For Frege, we understand thesense of a complex expression by understanding thesenses of its constituents» (1973, p. 152). Thecorrespondence between the parts of sentences andthe parts of thoughts plays an epistemic role inFrege’s views: our ability to discern thecorrespondence between the structures of sentencesand thoughts allows us to apprehend new thoughts.The structure of the sentence expressing a thoughtfunctions as a representation of the structure of

19

the expressed thought and facilitates our grasp ofit:

[O]ur thinking is closely bound up with language andthereby with the world of the senses. … To be sure, wedistinguish the sentence as the expression of thethought from the thought itself. We know we can havevarious expressions of the same thought. Theconnection of a thought with one particular sentence isnot a necessary one: but that a thought of which we areconscious is connected in our mind with some sentenceor other is for us men necessary. But that does notlie in the nature of the thought but in our own nature.There is no contradiction in supposing there to existbeings that can grasp the same thoughts as we dowithout needing to clad it in a form that can beperceived by the senses. But still for us men there isthis necessity (1924/25; Eng. tr., p. 269).

Since Frege undoubtedly stated many versions ofthesis (A) and provided many relevant examples ofsentences expressing the same thought, which hediscussed in detail, it is hard to accept anyinterpretation that claims that Frege held theview that thoughts are internally and uniquelystructured. On the contrary, Frege denies thatthe thought determines the structure reflected inthe sentence expressing it: «The thought itselfdoes not yet determine what is to be regarded asthe subject. If we say ‘the subject of thisjudgment’, we do not designate anything definiteunless at the same time we indicate a definitekind of analysis; as a rule, we do this inconnection with a definite wording» (1892a; Eng.tr., p. 188). Hence, the wording of the sentence

20

expressing a thought determines the parts of thecontent just like the different analysis of oneand the same sentence determine differentfunctions and arguments14. For Frege, the apparentstructure in the internal composition of a thoughtis due to the particular wording that is used toexpress a thought: «We must notice […] that oneand the same thought can be split up in differentways and so can be seen as put together out of partsin different ways» (1906c; Eng. tr., pp. 201-202,my emphasis). Thus, the structure of thoughts isdependent on the linguistic expression of thethought and on our way of seeing or grasping thiscontent. This is how the same thought can beexpressed by different sentences and why when thisoccurs the thought gets to be analyzed indifferent ways. The partition of a thoughtdepends on how the thought is expressed; forFrege, language has a crucial role in ourapprehension of thoughts.

All of the above has been defended elsewhere.Here I would like to submit that by reflecting onhow to connect the first two theses with theContext Principle, or thesis (C), one can bothprovide support for the above solution of thefirst puzzle and begin to see how to diffuse theapparent conflict between the last two theses,

14 «The proposition that Cato killed Cato shows the same thing. Here, if we think of ‘Cato’ as replaceable at the first occurrence, then ‘killing Cato’ is the function. If wethink of ‘Cato’ as replaceable at the second occurrence, then‘being killed by Cato’ is the function. Finally, if we think‘Cato’ as replaceable at both occurrences, then ‘killing oneself’ is the function» (1879, Eng. tr., p. 126-127).

21

i.e., between compositionality and contextuality.Frege might have held both of them if they are notinterpreted as supporting semantic atomism andsemantic holism, respectively. Instead, it is usefulto remember that for Frege, the Context Principlewas inextricably linked to his rejection ofsemantic psychologism and that it was above all anepistemological recommendation.

Indeed, if we accept Frege’s claim that theparts of sentences can lead us to graspingstructured thoughts, i.e. thought that have partscorresponding to the parts of the sentencesexpressing them, it may seem that the only way todo this is if the senses of words are ideas insome minds. Michael Resnik defines the type ofsemantic psychologism that Frege rejected as «thedoctrine that the meaning of a word (for a person)is the idea or mental image associated with thatword and that the meaning of a sentence is thecomplex idea composed of the ideas associated withits words» (1976, p. 37). This type ofpsychologism provides an account of how we graspthoughts, i.e., by grasping the complex ideaformed by the individual ideas linked to the wordsof the sentence expressing the thought. But Fregestrongly opposed this psychologism because if themeaning of a word is the mental image belonging toan individual mind, that meaning is private andsubjective:

The meaning and sense of a sign are to bedistinguished from the associated idea. If what a signmeans is an object perceivable by the senses, my ideaof it is an internal image, arising from memories of

22

sense impressions which I have had and acts, bothinternal and external, which I have performed. Such anidea is often imbued with feeling; the clarity of itsseparate parts varies and oscillates. The same sense isnot always connected, even in the same man, with thesame idea. The idea is subjective: one man’s idea isnot that of another. There result, as a matter ofcourse, a variety of differences in the ideasassociated with the same sense. … This constitutes anessential distinction between the idea and sign’ssense, which may be the common property of many people,and so is not a part or a mode of the individual mind.For one can hardly deny that mankind has a common storeof thoughts which is transmitted from one generation toanother (1892a; Eng. tr., p. 159-160)

A private mental content cannot be the meaningof a word because no one can ever access thecontent of someone else’s mind. If a mental imagewere the meaning of a word, we could never knowwhat meaning another person ascribes to a word;the meaning is unreachable for anyone but thesubject. Second, if the meaning of a word were amental image, there could never be any objectivecontent for a sentence; there could be neitheragreement nor disagreement between differentindividuals, since nothing would be publiclysharable. There could be no «common store ofthoughts» that generations after generations havetransmitted.

By keeping in mind Frege’s rejection of semanticpsychologism, we can appreciate how the ContextPrinciple actually fits with Frege’s views on theepistemic role of language in our apprehension ofthoughts. Frege needed a thesis like the Context

23

Principle to reject semantic psychologism, or theidea that the meaning of a word might beidentified with an idea in the mind; for semanticpsychologism would have seemed to many as theeasiest way to explain how we grasp the thoughtexpressed by a sentence. Yet, if we cannot learnthe meaning of a word by merely having an ideacorresponding to the word in our mind and if welearn words by learning the meaning of thesentences in which these words are used, then wecan maintain that there is a correspondencebetween parts of thoughts and parts of thesentences expressing these thoughts withoutfalling into semantic psychologism.

Frege’s acceptance of (B), (C), and (D) isconsistent. There is mutual reinforcement betweengrasping a thought on the basis of itsconstruction from its components and grasping themeaning of a word on the basis of learning itsrole in the sentences in which it occurs. In thefollowing passage, for example, Frege denies thatforming an idea of the content of a word isnecessary for its possessing a meaning: «That wecan form no idea of its content is therefore noreason for denying all meaning to a word, or forexcluding it from our vocabulary» (1884; Eng. tr.,§60). The rejection of semantic psychologismsupports semantic holism as an alternativeaccount. If the sense of a complex expression isformed by the composition of the senses of theparts and we can learn the meaning of these partsonly by learning what these words mean in othersemantic contexts, then compositionality andcontextuality work together to build an account of

24

how human learn the meaning of words that rejectssemantic psychologism and atomism:

We next laid down the fundamental principle that wemust never try to define the meaning of a word inisolation, but only at it is used in the context of aproposition; only by adhering to this can we, as Ibelieve, avoid a physical view of number withoutslipping into a psychological view of it (1884; Eng.tr., §106).

Frege’s concern with a subjective and thusincommunicable account of the meaning of (number)words led him to accept the Context Principle;this principle also allowed him to hold on to thethesis that the correspondence between thecomponents of sentences and the components ofthoughts is epistemologically crucial in theexplanation of how humans grasp thoughts withoutever having to resort to semantic psychologism.

10. Conclusion

We cannot claim with certainty that Frege sawany conflict between these four theses or that heexplicitly spelled out a solution to them. My moremodest suggestion is that these theses in Fregehave a common denominator, which is anepistemological concern to explain both howlanguage expresses thought and how humans grasplinguistic expressions of thoughts. I do not claimthat this was Frege’s solution to the allegedconflicts among these theses; I only submit that

25

this solution is Fregean in the sense of agreeingwith some deep concerns of Frege. Had Fregethought explicitly and deliberately about thesealleged conflicts, he might have developed asolution along the lines I propose.

References Angelelli I. (1967). Studies on Gottlob Frege and Traditional Philosophy.

Dordrecht: Reidel. Bell D. (1981). The Place of the Grudlagen in Frege’s

Development. The Philosophical Quarterly, 31, 124: 209-224. Bell D. (1987). Thoughts. Notre Dame Journal of Formal Logic, 28: 36-

54.Bell D. (1996). The Formation of Concepts and the Structure

of Thoughts. Philosophy and Phenomenological Research, 56: 583-596.

Bermúdez J.M. (2001). Frege on Thoughts and Their Structure.Logical Analysis and History of Philosophy, 4: 87-105.

Bronzo S. (manuscript). Frege on Multiple Analyses and theEssential Articulatedness of Thought.

Currie G. (1985). The Analysis of Thoughts. Australasian Journal ofPhilosophy, 63, 3: 283-298.

Dummett M. (1973). Frege. Philosophy of Language. London: Duckworth.Dummett M. (1975). What is a Theory of Meaning? In:

Guttenplan S., ed. Mind and Language. Oxford: OxfordUniversity Press: 97-138.

Dummett M. (1981). The Interpretation of Frege’s Philosophy. London:Duckworth.

Fodor J. and Lepore E. (1992). Holism. A Shopper’s Guide. Oxford:Oxford University Press.

Frege G. (1879). Begriffsschrift, eine der arithmetischen nachgebildeteFormelsprache des reinen Denkens, Nerbert, Halle (Engl. tr.:Conceptual Notation and Related Articles. Oxford: Clarendon Press,1972).

Frege G. (1880/81). Booles rechnende Logik und dieBegriffsschrift (En tr.: Boole’s logical Calculus and theConcept-script. In: Posthumous Writings. Oxford: BasilBlackwell 1979: 9-46).

26

Frege G. (1884). Die Grundlagen der Arithmetik - Eine logisch-mathematischeUntersuchung uber den Begriff der Zahl. Breslavia: Koebner (Engl.tr.: The Foundations of Arithmetic. Oxford: Blackwell, 1950).

Frege G. (1892a). Über Sinn und Bedeutung. Zeitschrift für Philosophieund philosophische Kritik 100: 25-50 (Engl.tr.: On Sense andMeaning. In: Collected Papers on Mathematics, Logic and Philosophy.Oxford: Basil Blackwell, 1984: 156-194).

Frege G. (1892b). Über Begriff und Gegenstand. Vierteljahrsschriftfür wissenschaftliche Philosophie 16: 192-205 (Engl.tr.: On Conceptand Object. In: Collected Papers on Mathematics, Logic and Philosophy.Oxford: Basil Blackwell, 1984: 182-194).

Frege G. (1893-1903). Grundgesetze der Arithmetik, begriffsschriftlichabgeleitet. Vol. 1, 1893, vol. 2, 1903, Pohle, Jena (Engl.tr.: The Basic Laws of Arithmetic. Exposition of the System. Berkeley andLos Angeles: University of California Press, 1964).

Frege G. (1899-1906?). Über Euklidische Geometrie. (Engl.tr.: On Euclidean Geometry. In: Posthumous Writings. Oxford:Basil Blackwell, 1979: 167-169).

Frege G. (1906a) Einleitung in die Logik (Engl. tr.:Introduction to Logic. In: Posthumous Writing. Oxford: BasilBlackwell, 1979: 185-196).

Frege G. (1906b) Über die Grundlagen der Geometrie. II.Jahresbericht der Deutschen Mathematiker-Vereiningung 15: 377-403(Engl. tr.: On the Foundations of Geometry: Second Series.In: Collected Papers on Mathematics, Logic and Philosophy. Oxford:Basil Blackwell, 1984: 293-340).

Frege G. (1906c). Kurze Übersicht meiner logischen Lehren(Engl. tr.: Introduction to Logic. In Posthumous Writings.Oxford: Basil Blackwell, 1979: 197-202).

Frege G. (1914). Logik in der Mathematik. (Engl. tr.: Logicin Mathematics. In: Posthumous Writings, Oxford: BasilBlackwell, 1979: 203-250.

Frege G. (1918-1919). Die Verneinung. Eine LogischeUntersuchung. Beitrage zur Philosophie des deutschen Idealismus, I:143-157 (Engl. tr.: Negation. In: Logical Investigations. NewHaven: Yale University Press, 1977: 31-53).

Frege G. (1919). Aufzeichnungen fur Ludwig Darmstaedter.(Engl. tr.: Notes for Ludwig Darmstaedter. In: PosthumousWritings. Oxford: Basil Blackwell, 1979: 253-257.

Frege G. (1923) Gedankengefüge, Beitrage zur Philosophie des deutschenIdealismus, III: 36-51 (Engl. tr.: Compound Thoughts. In:

27

Logical Investigations. New Haven: Yale University Press, 1977:55-77).

Frege G. (1924-1925). Erkenntnisquellen der Mathematik undder mathematischen Naturwissenschaften. (Engl. tr.: Sourcesof Knowledge of Mathematics and the mathematical naturalSciences. In: Posthumous Writings. Oxford: Basil Blackwell,1979: 267-274.

Frege G. (1976) Nachgelassene Schriften und wissenschavtlicherGriefweschsel, vol. 2, Felix Meiner ed., Hamburg (Eng. tr.:Gabriel G, Hermes H., Kambartel F., Thiel C. and VeraartA., eds. Philosophical and Mathematical Correspondence Chicago:University of Chicago Press 1980.

Garavaso P. (1991). Frege and the Analysis of Thougths. Historyand Philosophy of Logic, 12: 195-210.

Haaparanta L. (1985). Frege’s Context Principle. Communicationand Cognition, 18: 81-94.

Heck R.G. and May R. (2011). The Composition of Thoughts.Nous, 45, 1: 126-166.

Hodes H. (1982). The Composition of Fregean Thoughts.Philosophical Studies, 41: 161-178.

Janssen T.M.V. (2001). Frege, Contextuality andCompositionality. Journal of Logic, Language, and Information, 10:115-136.

Kemmerling A. (2011). Thoughts Without Parts: Frege’sDoctrine. Grazer Philosophische Studien 82(1): 165-188.

Künne W. (2007). A Dilemma in Frege’s Philosophy of Thoughtand Language. Rivista di estetica, 34, 1: 95-120.

Milne P. (1986). Frege’s Context Principle. Mind, 95, 380:481-495.

Pelletier F.J. (2001). Did Frege Believe Frege’s Principle?Journal of Logic, Language, and Information, 10: 87-114.

Penco C. (2003). Frege: Two Theses, Two Senses. History andPhilosophy of Logic, 24, 2: 87-109.

Resnik M. (1976). Frege’s Context Principle Revisited. In:Schirn M., ed. Studies on Frege I. Stuttgart-Bad Cannstatt:Frommann-Holzboog.

Ruffino M.A. (1991). Context Principle, Fruitfulness of Logicand the Cognitive Value of Arithmetic in Frege. History andPhilosophy of Logic, 12: 185-194.

Stuhlmann-Laeisz R. (2001). The Context-Principle. In: NewenA., Northmann U. And Stuhlmann-Laeisz R., eds. Building on

28

Frege New Essays on Sense, Content, and Concept. Stanford California:CSLI: 250-264.

Szabó Z.G. (2012) Compositionality. Stanford Encyclopedia ofPhilosophy,http://plato.stanford.edu/entries/compositionality/, lastaccessed May 20, 2013.

Textor M. (2009). A Repair of Frege’s Theory of Thoughts.Synthese, 167: 105-123.

29