From the WAB archives: A selection from the Bergen Wittgenstein archives' Working Papers and audio-visual materials

1.1 Material held at The Bertrand Russell Archives

The Bertrand Russell Archives at McMaster University, Canada, has three items relating to the Russell Version of "Notes on Logic". Item RA1.710.057822 is a typescript of 7ff. bearing handwritten corrections and additions, apparently by both Russell and Wittgenstein. It corresponds approximately to the first section of NL 1979 headed "Summary". Item RA1.710.057823 is a manuscript of 23ff.+iii in Russell's hand corresponding approximately to the remainder of the text of NL 1979, from the section headed "First MS" to the end. Item RA1.710.057824 is a typescript of two parts (i+8ff.+25ff.), the first part corresponding approximately to the contents of item RA1.710.057822 incorporating the handwritten corrections, and the second part corresponding approximately to the contents of RA1.710.057823.

Item RA824³ is foliated in two parts; i+1-8 and 1-25, and so may have been typed on two separate occasions. The whole item contains marginal handwritten Roman numerals which correspond to the enumerated headings used by Russell in his rearrangement of Wittgenstein's material, subsequently published as NL 1957, the Costello Version. At the end of RA823 there are three unnumbered folios, the first of which bears a contents page for the Costello Version. It would appear that this contents page was used as a guide for the annotation of RA824 and its subsequent rearrangement. The final two unnumbered folios at the end of RA823 each contain one diagram. The first is not contained in NL 1979, but is similar in arrangement, though not in content, to that found in AM 1979 p.115. The second has no counterpart in TB 1979.

The catalogue numbering of the three items in the Russell Archives corresponds to their chronological sequence. Two types of evidence may be found for this assertion. Firstly, in amendments to the text which have been incorporated into subsequent scripts. For example: RA822, f.1, ¶5 reads

The ^One reason for supposing that not all
propositions which have more than one
argument are relational propositions is that
┌ ^if ┐ they were‹,› the relations of judgement
and inference that ^would have to hold
between an arbitrary number of things.⁴

Typescript RA824, Part 1, f.1, ¶5 reads

One reason for supposing that not all
propositions which have more than one
argument are relational propositions is
that if they were, the relations of
judgment and inference would have to hold
between an arbitrary number of things.

Similar evidence may be found of corrections in RA823 which have been incorporated into the second part of RA824, for example: RA823, f.11, ¶1 reads

At a pinch, we are always inclined to
explanations of ┌ ^{logical ┐} functions of
propositions which ┌ ^{aim at introducing into
the function ┐} either only contain the
constituents of these propositions, or only
their forms, etc. etc; & we overlook that
ordinary language would not contain the
whole propositions if it did not need
them:...

Typescript RA824, Part 2, f.12, ¶2 reads

At a pinch we are always inclined to
explanations of logical functions of
propositions which aim at introducing into
the function either only the constituents of
these propositions, or only their form, etc.
etc.; and we overlook that ordinary language
would not contain the whole propositions if
it did not need them: ...

All ampersands in the manuscript are spelled out fully in the typescript as "and". Russell invariably used the ampersand in these scripts, except at the beginning of a sentence.

As there is no overlap of the contents of RA822 and RA823, it is not possible to use such evidence to determine the chronology of these two items. However, a study of the sequence of events surrounding the provenance of the two scripts, coupled with certain grammatical observations, may serve to confirm the assertion above, that RA822 precedes RA823.

Wittgenstein worked on "Notes on Logic" in Norway while on holiday with David Pinsent from 30 August 1913 to 1 October 1913. On his return he went first to Cambridge, on 2 October; and then to Birmingham, on 6 October, to stay with Pinsent's family. On 7 October, he dictated some material to a German-speaking typist at the Berlitz School of Translation in Birmingham. On 8 or 9 October he returned again to Cambridge where he was in conversation with Russell about his ideas⁵. Russell had the conversations recorded in shorthand by Jourdain's secretary⁶. He may have taken the Berlitz typescript to Cambridge with him and given a copy of it to Russell. However, the need for a dictation to Russell suggests, on the contrary, that Russell did not have a written account of Wittgenstein's ideas before him. Wittgenstein departed Cambridge for London and thence, on 11 October, to Norway once again, via Newcastle. He stayed in Norway until the end of June or the beginning of July 1914, breaking this period only with a brief visit to Vienna at Christmas⁷.

Von Wright (1974 p.29) describes the shorthand notes taken by Jourdain's secretary as a "summary". Russell had a typescript prepared from these notes, which appears to be item RA822 and is entitled "Summary". Corroborating of this identification, von Wright states that the typescript contains a misprint of "polarity" for "bi-polarity", mentioned by Wittgenstein in a letter to Russell (listed as R20)⁸ dated November 1913. Accordingly, item RA822, p.2, ¶4 reads

To understand a proposition p it is not
enough to know that ”p implies ‹'”›p” is
true‹'›, but we must also know that p also
implies ‹'› ”not‑p” is false ‹'› ^{~p implies
”p is false”}. This shows the ┌^bi┐polarity
of the proposition.

There appear to be corrections in the hand of both Russell and Wittgenstein. This would suggest that typescript RA822 was sent to Wittgenstein in Norway after it had been prepared from the shorthand notes. Wittgenstein made various changes to the text, and returned it to Russell who made further corrections and had item RA824 prepared from it.

The typescript Wittgenstein had prepared at the Berlitz School for Russell was in German. Item RA823 is in Russell's hand, and appears to be a translation of material by Wittgenstein. The reason for supposing it to be a translation is that in places the English is broken, but in a way which would correspond to German-English transliteration. For example, item RA823 f.12 ¶1 reads

Just as little as we are concerned, in
logic, with the relation of a name to its
meaning, just so little are we concerned
with the relation of a proposition to
reality,...

The German original is no longer extant, but the translation in W1 1984 p.200, which is a translation from the above English back into German, reads

Ebensowenig, wie wir uns in der Logik für die Beziehung eines Namens zu seiner Bedeutung interessieren, sowenig interesseiren wir uns für die Beziehung eines Satzes zur Wirklichkeit...

Von Wright asserts (1974 p.30) that item RA823 is the translation of Wittgenstein's Berlitz typescript. This opinion is also asserted by McGuinness (1972 p.448). However, in the letter to which von Wright refers (R19), Wittgenstein asks of Russell "Did you get the copy of my manuscript?" The reference appears to be to the Berlitz material. If this does refer to the typescript prepared at the Berlitz School then it seems that Wittgenstein did not take it to Russell, but that it was forwarded to Russell from Berlitz after Wittgenstein had finally left for Norway. This accords with the earlier observation that Russell appeared not to have a written account before him (p.7 above). As Wittgenstein last saw the material in manuscript form this may explain why he mistakenly refers at this point to a manuscript rather than a typescript. Alternatively, McGuinness believes the above indicates that Wittgenstein sent Russell the original German manuscript (1972 p.448).

The sequence of events may now be summarised:

Wittgenstein worked on NL on holiday in Norway in 1913. He returned, and commissioned a typescript in German from Berlitz which was generated via some shorthand notes. Meanwhile he went to Cambridge without the typescript and dictated a summary to Russell in English which was taken down in shorthand. Russell had this summary typed (item RA822). Wittgenstein departed for Norway, meanwhile the German typescript arrived to Russell from Berlitz. Russell translated it (item RA823). Russell sent a copy of item RA822 to Wittgenstein in Norway, who amended it and returned it to Russell. Russell added the amended item RA822 (the ad-hoc dictation) to the translation item RA823 (from Wittgenstein's prepared German dictation) and had them retyped as item RA824.

After this, around February 1914⁹, Russell made a rearrangement which is indicated by the marginal Roman enumeration of item RA824 and which follows the contents list on one of the unnumbered folios at the end of RA823. This rearrangement, possibly still in manuscript, was in his possession when he went to Harvard on 7 March 1914¹⁰ where it was copied by Costello during the three weeks he was Russell's assistant. Costello states that Russell's manuscript was dated September 1913 (NL 1957 p.230), but Russell must have made the arrangement later.

1.2 Inferred Material now Lost

It has been shown above that originally there were some manuscripts and typescripts which have since been lost. These scripts can be inferred from the writings of the parties involved, e.g. Pinsent's diaries. Other manuscripts may be inferred, even if they are not explicitly referred to, e.g. the four "manuscripts" used by Wittgenstein for the Berlitz dictation as implied by the sub-titles. In addition, some commentators have mentioned material which seems to be no longer extant, e.g. additional notebooks to the three still existing, mentioned by Engelmann in a letter to Hayek dated 23 April 1953 (cited by McGuinness et al 1971 p.4; see also McGuinness 1988 p.187 footnote 7 "My conjecture is that at least one of these notebooks contained Notes on Logic material").

McGuinness (1989 p.37f.) describes a list of material in Hermine Wittgenstein's handwriting which appears to confirm the earlier existence of the lost notebook mentioned above (referred to as item 1).

The following list has been compiled with reference to the texts cited, or inferred from them:

1.3 TS_x

The Wittgenstein Archives at the University of Bergen has a photocopy of a TS of "Notes on Logic" in the Russell Version. It was obtained by Alois Pichler of the Archives, from the collection of G.H. von Wright, in 1990. However, it contains a different number of pages to that stated in von Wright's catalogue under 201a or 201b.

I will call the Archives' typescript TS_x for the time being. TS_x has 15 folios, typed on one side of the folio only¹¹, and foliated 2-14 preceded by 2 unnumbered folios. The size of the folios as recorded on the A3 photocopy is 203mm x 329mm (approximately equal to UK Foolscap or US Legal). The photocopy appears to be 1:1 to the original as assessed by the typesize which is 10 characters per inch, an imperial standard. The text is arranged under five headings: "SUMMARY" (on unfoliated f.ii), "First MS" (on f.4), "Second MS" (on f.5), "Third MS" (on f.7) and "Fourth MS" (on f.11). TS_x has typescript pagination in the format "- 1 -" which, on the evidence of alignment with the body copy, was made at the same time as the typing of the text¹².

There are two diagrams inserted by hand in TS_x. The first is on f.12 and shows A standing in relation a-p-b to the poles of the proposition a and b. This diagram was published in NL 1979 (p.106), but not in NL 1957. The text of TS_x ends about one third of the way down f.14 after which there occurs a second diagram. This diagram is the same as that on the second unnumbered folio f.25 of RA823.

1.4 Von Wright 1969 and 1982

The first edition of the von Wright catalogue was published in 1969, prior to McGuinness' researches. It describes the TSS of "Notes on Logic" in the following way:

[p.492] 201a "Notes on Logic" September 1913. The so-called Russell Version. English.

201b The same. The so-called Costello Version. English.

[p.498] The history of these notes is obscure. There exist two versions of them, both in English [footnote: See the editorial note in Notebooks 1914-1916, p.93]. Both are dated September 1913 [footnote: See Mr. Costello's introductory remarks to the publication of 201b in the Journal of Philosophy, 54, (1957), 230f.]. From Wittgenstein's letters to Russell one gets the impression that Russell got them from Wittgenstein in October. There are several indications that they originally existed in manuscript form and that Russell had the manuscripts typed and sent it to Wittgenstein, who was then in Norway [footnote: See Notebooks 1914-1916, p.123]. Perhaps Wittgenstein then revised this typescript and returned it to Russell or perhaps he sent Russell a new manuscript. In any case, 201b appears to be a revised version of 201a, the revisions being made by Wittgenstein himself.

A revised edition of the von Wright catalogue was published in 1982, following the publication of McGuinness' article. It describes the TSS in the following way:

[p.46f.] 201a "Notes on Logic" September 1913. The so-called Russell Version. English. 7 typescript pages dictated by Wittgenstein and 23 manuscript pages in Russell's hand

201b The same. The so-called Costello Version. English.

[p.54] There exist two versions of these notes. Both are in English and both date from the autumn of 1913. Their origin and mutual relation were for a long time obscure, but have eventually been clarified in what seems a conclusive manner by Brian F. McGuinness in "Bertrand Russell and Ludwig Wittgenstein's "Notes on Logic "", Revue Internationale de Philosophie 26, 1972. What is called the Russell Version in the catalogue consists of a "Summary" evidently dictated by Wittgenstein (in English) and four "Manuscripts" which Russell had translated into English from notes in German by Wittgenstein. The so-called Costello Version is apparently a subsequent rearrangement of the text made by Russell alone.

Several matters arise from this description by von Wright. First, that TS_x does not fit either description. Second, the description of the Costello Version as "the same" cannot be accurate. Assuming that 201b had the headings that were published in NL 1957, then 201b was arranged under seven headings; a "Preliminary" followed by six Roman enumerated headings with titles. On the other hand, 201a is described above as having five headings; a "Summary" followed by four untitled "manuscripts". It is possible that von Wright means by "the same", the same number of pages, or the same number of manuscript and typescript pages, or the same combination of both manuscript and typescript pages. Unfortunately, none of these possible variants for the description of 201b, nor the description of 201a, fits TS_x which contains 16 typescript folios altogether and with no manuscript folios.

1.5 Anscombe and von Wright 1960 and 1979

There is an editorial preface to NL 1960 which reads as follows:

Two versions of these notes exist, one in the possession of Lord Russell and the other published by Mr. H.T. Costello in the Journal of Philosophy for April 25th, 1957, Volume LIV, No. 9. Mr. Costello got his version from Lord Russell in 1914. Lord Russell himself is unable to remember how there came to be two versions. We are indebted to Mr. J. Griffin, of St. Anthony's College, Oxford, for bringing it to our attention that there is convincing internal evidence to show that Mr. Costello's version was made by Wittgenstein himself, being an arrangement, with some small alterations, of the earlier material. Since this version seems to be an improvement on the earlier one, we reproduce it here, with some small corrections of copyist's errors. [Edd.]

In the preface to NL 1979, the editors von Wright and Anscombe mention the 1972 publication of McGuinness. In it they write

It was clear that the Costello<e> version was a slightly corrected total rearrangement of that text under headings... the Costello<e> version was constructed by Russell. The other one is therefore closer to Wittgenstein, the first part of it being his own dictation in English and the rest a translation by Russell of material dictated by Wittgenstein in German.

1.6 McGuinness 1972

McGuinness describes a number of copies and versions of the material called "Notes on Logic". However, his description is difficult to follow because he does not give names or specific references to all of the items to which he refers. I shall make a brief summary together with source page references to his article in square brackets [ ].

NL 1957 was published from a copy, made by Costello, of a "manuscript" in the possession of Russell when he visited Harvard in March 1914. Both the original and the copy are now lost [p.444].

Two earlier stages exist and are in the collection of The Bertrand Russell Archives at McMaster University [p.444].

Stage 1 consists of two parts. First, a typescript of seven sheets with manuscript corrections by Russell and Wittgenstein and the title "Summary" in Russell's hand. The second, a manuscript of 23 sheets in Russell's hand entitled "Wittgenstein" on f.1, and bearing the sub-headings "First MS" on f.1, "2nd MS" on f.4, "3rd MS" on f.7, and "4th MS" on f.16 [p.445]. (I shall use the McMaster references and call the former RA822 and the latter RA823).

Stage 2 consists of two typescripts, one foliated 1-8 and entitled "Summary". The second foliated 1-25 containing the following headings: "First MS" on f.1, "Second MS" on f.4, "Third MS" on f.8, and "Fourth MS" on f.17. (these together, following McMaster, I shall call RA824 although the total number of folios in the first typescript is nine: i+8ff.) [p.444].

McGuinness does not identify which TS or set of MSS and TSS is the one named 201a in von Wright's 1969 catalogue. Von Wright's description matches the pair of scripts in Stage 1: RA822 and RA823. However, it is the pair of typescripts in Stage 2, RA824, which McGuinness says "are evidently those shown by Russell to Mr. D. Schwayder" [p.444].

1.7 The Identity of Sources

From §1.6 above, several questions arise: (1) is 201a identical with items RA822 plus RA823?; (2) are items RA822 plus RA823 identical with RA824?; (3) how is TS_x related to these items?; (4) which source seems to be the one from which NL 1979 was produced?; (5) does this affect the literature regarding the provenance of NL 1979?

(1) Von Wright (1982) describes 201a as "7 typescript pages dictated by Wittgenstein and 23 manuscript pages in Russell's hand." This shows that 201a is identical with the two scripts RA822 and RA823 combined.

(2) Item RA824 is a typescript from RA822 and RA823 in which instructions for deletions and additions are carried out. Therefore RA824 is not identical with RA822 and RA823. If the distinction between instruction and execution of deletions and additions is ignored, there still remain the following differences: RA824 contains further deletions at the following points: f.2 (part 1) ¶1; f.4 (part 1) ¶1; f.5 (part 1) ¶1; f.3 (part 2) ¶4; f.19 (part 2) ¶6; f.20 (part 2) ¶2 and f.21 (part 2) ¶3. Apart from the necessary insertion of non-standard typewriter characters, RA824 also includes manuscript insertions at the following points: f.6 (part 1) ¶3 and f.20 (part 2) ¶2. The diagrams in RA823 ff.25 & 26 are omitted from RA824.

(3) Deletions in RA824 may be compared to the text of TS_x with a variety of results. The following deletions in RA824 are carried out in TS_x: f.2 (part 1) ¶1; f.21 (part 2) ¶3 and f.20 (part 2) ¶2. The following deletion is copied as text which is then deleted: f.5 (part 1) ¶1. The following deletions in RA824 are ignored and the text included in TS_x: f.2 (part 2) ¶4 and f.19 (part 2) ¶6, but with the supplementary comment "This was typed in but had exesses through it (D.S.)". Finally, the diagram in TS_x f.14 does not appear in RA824 but derives from RA823 f.25 with an additional annotation. It therefore cannot be determined from this evidence whether TS_x derives directly from RA824.

The last mentioned paragraph in TS_x (f.19 (part 2) ¶6), and its comment, suggest that TS_x was copied from another typescript and not from RA824, which does not contain the deletion in the form described. In addition, the comment shows it is unlikely to have been copied from the manuscript RA823 owing to the style of the original deletion. Furthermore, the deletion of this paragraph first appears in the later RA824. It would therefore seem that TS_x was copied from an unknown TS which derived from RA824 and adopted a mixed approach to the inclusion of handwritten comments.

(4) Following (2) above, a comparison was made between the content of the McMaster material and NL 1979. On the basis of the evidence cited, it was concluded that NL 1979 is closer to the text of RA824 than to the combined texts of RA822 and RA823.

NL 1979 includes references to the Tractatus which also occur in TS_x. Furthermore the contents of TS_x are close to that of NL 1979 (a full analysis is appended in Part II beginning on p.61). In particular, NL 1979 follows the omission of the negation sign "~" in TS_x f.2, footnote *, which reads "the old poles are correlated to p". RA824 reads "the old poles are correlated to ~p", as does RA822.

Conversely, TS_x f.3 ¶3 reads "to the usual indefinables" whereas NL 1979 reads "to old indefinables". RA824 reads "to /the usual/ old indefinables". RA822 reads as NL 1979.

From this it is not possible to say whether NL 1979 is closer to RA824 than to TS_x.

(5) The editors state in NL 1979 that they used the Russell Version for the preparation of the publication. Von Wright 1982 states that 201a is the Russell Version. From his description it follows that 201a refers to RA822 and RA823. However we have seem from (1) and (4) above that the editors also had recourse to RA824. In addition, we have seen in (4) above that some features of TS_x which are not features of any of the McMaster material are also included. From this it may be concluded that the editors made use of a variety of sources in the publication of NL 1979, and not only the von Wright catalogue item 201a (RA822 and RA823) as implied in the Preface. It may also be objected that in giving these two items a single identity, von Wright's catalogue is somewhat misleading.

1.8 Diagrams

There is one diagram in NL 1979 (p.106) but this same diagram is omitted from NL 1957. As TS 201b, the source corresponding to the latter, is missing it is not known whether or not the diagram was included by Russell. However, the sources of NL 1979 reveal two further diagrams. They are in RA823 on folios 25 and 26 respectively, though they may be regarded as loose sheets appended to the original composition at a later date.

The diagram in RA823 f.25 shows Wittgenstein's bi-polar diagrammatic representation of two propositions. At this time his notation for these poles was "a" and "b". Russell's marginal note in RA822 (f.2) indicates that the "ab" notation corresponds to the later "TF" [WF] or "True-False" [Wahr-Falsch] notation.

The function is not specified but, assuming a=T and b=F, the following table which more clearly shows the truth conditions may be constructed:

This is the truth table, and therefore the diagrammatic representation, of p│q or ~p∨~q. This function is mentioned in RA823 on folios 13f. and it may be that the diagram is a graphical representation of the text at that place in the manuscript. The same diagram appears on the last folio of TS_x (f.14) with the handwritten annotation "This is the symbol for ~p∨~q". Although the annotation is correct it remains curious that the diagram remains isolated from the text to which it may refer.

The diagram may be compared with other similar ones which Wittgenstein used at this time. Immediately after arriving in Norway, and as a commentary on the material sent to him by Russell, Wittgenstein replied with a letter which included a diagram (dated Norway 1913). The diagram was infelicitously transcribed for its publication in S1 1960 (p.272) and TB 1961 (p.126).

It was corrected for its publication in TB 1979 (p.127). The Preface mistakenly refers to this correction occurring on p.126:

It is clear from the text of Wittgenstein's letter (NL 1979 p.129) that Russell did not understand the notation he was using:

I am upset that you did not understand the rule for the signs in my last letter... This is the sign for p≡p; it is tautological because b is connected only with such pairs of poles as consist of opposed poles of a proposition (p); if you apply this to propositions with more than 2 arguments, you get the general rule according to which tautologies are constructed.

It is therefore possible that Russell infelicitously transcribed the diagram at RA823 f.25 from this letter in an attempt to understand Wittgenstein's notation. An example of a similar misunderstanding may be found in Moore's notes taken as dictation from Wittgenstein during his stay with him in Norway in 1914. Wittgenstein dictated the method of showing a tautology to Moore. This diagram was felicitously transcribed from Moore's notebooks for its publication in AM 1960. However, Moore's original diagram in D 301 does not show the truth combinations of tautology and it remains uncorrected or annotated in all subsequent editions¹³. Iglesias (1981 p.318) is alone in commenting on this disjunction between graphics and text.

Using the truth table notation for tautology the requisite relations may be shown clearly:

However, the truth table corresponding to Moore's diagram is as follows:

An alternative source for Russell's diagram in RA823 f.25 may therefore be Moore's notes in von Wright catalogue item 301, which were shown to Russell before the end of July 1914¹⁴. There is no evidence that the diagram in RA823 must have been present in the material taken to Harvard by Russell in March of that year. Russell's diagram is identical to Moore's and both attempt to show tautology, the former as p≡p and the latter as p≡~(~p).

Wittgenstein himself found this graphical method confusing. In the above mentioned letter to Russell he anticipates its refinement

even if this notation should turn out not to be the final correct notation...

One development was the move from "ab" to "TF" [WF] notation, but still the graphical method proved less clear than the truth tables which were derived from it. In the later TS 202 dating from around 1918 he includes a number of handwritten diagrams in the inserted section 6.1203. The second diagram shows Wittgenstein's original notation corrected in red ink:

The correction was made between the publication of TLP 1921 §6.1203, in which the original annotation was reproduced:

and the publication of TLP 1922 §6.1203 in which the diagram was corrected:

It is interesting to note that the correction has been made to typescript TS 202, used for the 1921 edition, since von Wright (in Wittgenstein 1973 p.vii) asserts that the source of the 1922 edition is a corrected off-print of the 1921 edition.

Black (1964a p.323) notes that the rules for the use of the graphical demonstration require an additional instruction in order to operate correctly:

forbidding the drawing of lines between the positive and negative poles of the same proposition.

1.9 Argument regarding TS_x

TS_x contains handwritten references, in an unknown hand, to the Tractatus numbering system of 1918. If the typescript was copied directly from Russell's manuscript then the handwritten references must have been added later. A distinctive feature of NL 1957 is that it does not contain references to the Tractatus. This would be as one would expect if Costello had simply copied the material in 1914 at Harvard.

TS_x has typescript comments embedded in the text and initialled "D.S.". This probably refers to D. Schwayder. McGuinness (1972 p.444) mentions

These typescripts [item RA824] are evidently those shown by Russell to Mr. D. Schwayder in the early 1950's, at which time copies made from them enjoyed a certain circulation.¹⁵

McGuinness therefore seems to suggest that D.S. copied the typescript RA824, integrating the two parts under a uniform pagination. If TS_x is the result then he also added some comments of his own on readings of the text. However, the text of TS_x also differs slightly in content from RA824, e.g. (3) above. This suggests that by the 1950's, Russell was using a different copy of the "Russell Version" which he showed to Schwayder. However, what is noteworthy is that the Costello Version does not contain this later text and so TS_x is unlikely to have been used as a basis for the much earlier Costello Version. On the contrary, there is no evidence in TS_x to support McGuinness' assertion that the Costello arrangement was made between 1950 and 1957 from the Schwayder copy, but rather that the arrangement was already made by Russell in 1914. It may be that McGuinness mistakenly writes "The Costello Version is obviously a rearrangement under chapter-headings of the Schwayder Version..." meaning "of the typescripts shown to Schwayder". However, the role Schwayder plays in this argument is unclear.

TS_x contains the footnote which reads "Russell for instance..." (f.ii). McGuinness, mentioning this footnote, writes:

The Costello Version is obviously a rearrangement under headings of the Schwayder Version. Some doublets are dropped, a sentence is added to compensate for a change of context, the English is improved, and some references to "you" or "I" (sic) are changed to references to "Russell". [p.444]

This suggests that the Schwayder Version does not read: "Russell...". However, TS_x does not fulfil this requirement.

1.10 Conclusion

The continued misrepresentation of the diagrams perpetuates the initial confusion about the bi-polarity of propositions. This confusion was later resolved by the introduction of truth-tables, but shows that Wittgenstein's own alternative representation was, and continues to be, undervalued.

McGuinness is incorrect when he writes that "the Costello Version is obviously a rearrangement under chapter-headings of the Schwayder Version" [p.444]. Costello is therefore correct in reporting that he copied "some notes and excerpts" in the possession of Russell at Harvard in 1914.

Some details of TS_x indicate that there was an intermediate typescript between RA824 and TS_x. If Costello made a copy from RA824 in 1914, this intermediate typescript was created between 1914 and 1950. It may have been a copy of the typescript which was cut up when Russell made his rearrangement (Costello Version). McGuinness (1972 p.453f.) describes the process of

cutting out the paragraphs, putting them in seven piles, arranging each pile in the desired order by shuffling (and in many cases by sub-dividing individual paragraphs), throwing away unwanted doublets, pasting the remaining slips on new sheets, and then writing in manuscript additions or corrections to improve the style and the continuity.

If an intermediate typescript was used for this purpose, this contradicts McGuinness' footnote 14 on p.454

This was presumably done with a duplicate of the typescript that we still possess.

Bearing in mind the number of intermediate scripts leading to the Russell Version as published in NL 1979, and bearing in mind that the role of these scripts as part of Wittgenstein's Nachlaß is negotiable; I propose that a new catalogue differentiation be created between the Russell Versions. Using von Wright's system of references as a basis, if the Russell Archive has 201a-1 (RA822+RA823) and 201a-2 (RA824), then TS_x can be called 201a-3. It is therefore also appropriate to refer to RA822 as Part 1 of 201a-1, and to RA823 as Part 2 of 201a-1. The intermediate typescript or manuscript referred to above seems now to be lost.

Using this terminology, NL 1979 appears to be closest to 201a-2 and 201a-3 (cf. §1.9(4) above). 201a-1 matches the physical description of item 201a in von Wright's present catalogue (cf. §1.9(1) above).

2.1 Comments on variations between NL 1957 and NL 1979

No detailed study of the relationship of the passages in the Costello and Russell Versions of "Notes on Logic" has hitherto been published¹⁶. In this Part, two lists are appended, each showing the correspondence of individual passages in the one to the other. The two lists are arranged in order of the passages in NL 1957 and NL 1979 respectively. The results of this analysis are summarised here.

Despite the different structure of these two versions, there is a very close correspondence between them. However, in the absence of a detailed comparison such as that presented below, earlier commentators have failed to correctly quantify the correspondence. For example Black (1964b p.133) claims

Everything in the second version [NL 1957] is to be found somewhere in the first, with the possible exception of the remark "It is wrong to conceive every proposition as expressing a relation".

NL 1979 was published from Wittgenstein's dictated (etc.) material as compiled by Russell. NL 1957 was published from Russell's later reorganisation of the material under new sub-headings. The contents of NL 1957 may therefore be considered as a derivative of the contents of NL 1979, despite the chronology of their publication. This is confirmed in the analysis below since NL 1957 does not contain any remarks which are not also in NL 1979. However, there are fourteen remarks in NL 1979 which do not occur in NL 1957. Some remarks which appear in both versions contain minor adjustments of grammar, e.g. NL 1957 p.236 ¶6 and NL 1979 p.102 ¶2; and some conditionals are removed, e.g. NL 1979 p.104 ¶6 "may be symbolized" to "symbolizes". There are also some, more major revisions, caused by rearrangement of the syntax. Other remarks are changed from a positive construction to a negative, e.g. NL 1979 p.102 ¶2 "Just as little as we are concerned..." to NL 1957 p.236 ¶6 "We are not concerned...".

Four corrections or variations introduce misreadings into the text:

1NL 1957 p.237 ¶3 misrepresents the expression of NL 1979 p.98 ¶10, changing it from:

Similarly in "φx" "φ" looks like a substantive but is not one; in "~p", "~" looks like "q" but is not like it.

to:

Similarly in "φx" "φ" looks like a substantive but is not one; in "~p", "~" looks like "φ" but is not like it.

This error was overlooked in McGuinness's "Appendix".

2NL 1957 removes the opening expression of NL 1979 p.104 ¶2 "If a word creates a world...", which McGuinness notes may allude to "the Word of God" (footnote 16, p.454).

3 The diagram in NL 1979 p.106 ¶3 is omitted.

4 McGuinness lists a number of occurrences where the original MS and TS material includes insertions by Russell in square brackets. In NL 1957 all of these are changed to round brackets, and square brackets are used once for the inclusion of what was a footnote into the body of the text (NL 1957 p.240 ¶1b, NL 1979 p.94 footnote 2). This confuses such insertions with Russell's in a manner criticised by McGuinness [p.458].

2.2 Comments on variations between NL 1957 and NL 1960

NL 1957 does not contain any references to the text of the Tractatus Logico-philosophicus whereas there are 38 such references in NL 1960. In addition to the editors' prefatory remarks, there are eight editorial footnotes to the English text of NL 1960, of which two relate to translation (pp.189 & 199), three clarify terminology (pp. 197, 203 & 215-2), one notes the editorial insertion of punctuation (p.223), one notes a variation between NL 1957 and Russell's copy of the typescript used for NL 1960 (p.201), and one notes an incongruous phrase (p.215-1). Of these, the footnotes on pp.189 & 199 have no counterpart in the German text.

The German text has eight footnotes of which two do not have counterparts in the English text. In the first of these the editors draw attention to some inserted text taken from Tractatus Logico-philosophicus §4.063 (p.192). In the second the editors mention an alternative expression used by Wittgenstein (p.204).

NL 1957 has one footnote by Costello (p.241) which appears as an editorial reference to the text in a footnote in NL 1960 (footnote 2, pp.214 & 215).

In addition to some minor variations of style, e.g. the inclusion of commas after i.e., and e.g.; and the hyphenation of "assertion-sign" there are thirteen occurrences of significant variation between the English text of NL 1957 and NL 1960.

1NL 1960 corrects an omitted hyphen and reads "A judges that p is true and not-p is false", restoring the opposed pole to the statement (p.195). There are four occurrences of improvements to the hyphenation of negatives (pp.195 & 207 (3)).

2NL 1960 includes a typographical error which reads "sence" (p.199) for "sense" (NL 1957 p.235).

3NL 1960 reads "Among the facts which make »p or q« true there are also facts which make »p or q« true" (p.199). NL 1957 correctly reads "Among the facts which make »p or q« true there are also facts which make »p and q« true" (p.236).

4NL 1960 reads "What symbolises..." (p.205). NL 1957 reads "What is symbolized... (p.237)

5 There is an error in NL 1960 in the English text alone (p.209) which reads "Signs of the forms »p∨~p« are senseless, but not the proposition »(p)q∨~p«.", whereas the German text, and that of NL 1957 read "Signs of the forms »p∨~p« are senseless, but not the proposition »(p)p∨~p«." (p.239).

6NL 1960 correctly reads "»~« looks like »_« but is not like it" (p.205) whereas NL 1957 reads ""~" looks like "q" but is not like it" (p.237).

7NL 1960 corrects an omitted letter and reads "If, for instance, »apb« says p, then bpa says nothing (it does not say ~p)" (p.211), whereas NL 1957 reads "If, for instance, "apb" says p, then bp says nothing (it does not say ~p)" (p.240).

8NL 1960 reads "And how do matters stand with the definition of »⊃« by »∨« and »~«, or of »∨« by »~« and »⊃«?" (p.213), whereas NL 1957 reads "And how do matters stand with the definition of »⊃« by »∨« and ».«, or of »∨« by ».« and »⊃«?" (p.240).

9NL 1960 correctly reads "arbitrary cases" (p.217) whereas NL 1957 reads "arbitrary causes" (p.241).

10NL 1960 incorrectly reads "p.p⊃q.⊃b'd.q" in the English text only (p.217). A reprographic error appears to have caused an inversion of the subscript element in the expression "p.p⊃q.⊃p,q.q" which is correctly expressed in the German text and in NL 1957 (p.241). The alignment of the text and the appearance of the comma suggest that the element was pasted-in upside down.

11NL 1960 omits the word "for" (p.219) from the sentence in NL 1957 which reads "We must not introduce it first for one class of cases..." (p.242).

12NL 1960 correctly records as "(∃x,y)xRy" (p.219) the expression in NL 1957 which reads "(∃xy).xRy" (p.242).

13NL 1960 reads "p̂[(∃x)._x=p]" (p.223) whereas NL 1957 reads "p̂(∃x)φx=p" (p.244).

Of these thirteen significant variations between the two editions, four result in improvements in the later edition and four in impairments.

2.3 Comments on variations between NL 1961 and earlier editions

NL 1961 is a corrected impression of the English content of NL 1960 and, excluding the adoption of the English typographic convention for quotation marks, varies from it in the following respects:

Correction of the error in NL 1960 p.199 (Cf. above)

Correction of the error in NL 1960 p.209 (Cf. above)

Correction of the error in NL 1960 p.217 (Cf. above)

2.4 Correspondences between NL 1957 and NL 1979

Table I is indexed by the pages of the Costello Version (NL 1957), table II by the pages of the Russell Version (NL 1979). These tables adopt the convention of numbering every paragraph on each page. Single paragraphs which run over more than one page therefore have more than one number. Parts of paragraphs are indicated by the suffixes a, b, c, etc.

2.5 Phrases in NL 1979 which are omitted from NL 1957

93	3	The verb of a proposition cannot be "is true" or "is false", but whatever is true must already contain the verb
93	4	Deductions only proceed according to the laws of deduction but these laws cannot justify the deduction
94	footnote 1	W-F = Wahr-Falsch-i.e. True-False
95	3	The form of a proposition has meaning in the following way. Consider a symbol "xRy". To symbols of this form correspond couples of things whose names are respectively "x" and "y". The things xy stand to one another in all sorts of relations, amongst others some stand in the relation R, and some not, just as I single out a particular thing by a particular name I single out all behaviours of the points x and y with respect to the relation R. I say that if an x stands in the relation R to a y the sign "xRy" is to be called true to the fact and otherwise false. This is a definition of sense.
95	5a	It is not strictly true to say that we understand a proposition p if we know that p is equivalent to "p is true" for this would be the case if accidentally both were true or false.
95	7	The ab notation makes it clear that not and or are dependent on one another and we can therefore not use them as simultaneous indefinables.
98	4	Not only must logic not deal with [particular] things, but just as little with relations and predicates.
98	8	Propositions are not names.
98	14b	...to understand it, we must know both what must be the case if it is true, and what must be the case if it is false.
101	7b	However, e.g., "not-p" may be explained, there must always be a meaning given to the question "what is denied?"
104	4	To understand a proposition means to know what is the case if it is true. Hence we can understand it without knowing if it is true.
105	6	It is very easy to forget that, though the propositions of a form can be either true of false, each one of these propositions can only be either true or false, not both.
106	3
106	12	Propositions can never be indefinables, for they are always complex. That also words like "ambulo" are complex appears in the fact that their root with a different termination gives a different sense.
107	3	Facts cannot be named.

2.6 Correspondences of page references between NL 1957, NL 1960 and NL 1961

McGuinness makes reference to the Costello Version published in NL 1961. As the page references differ between the three issues of the Costello Version a table of correspondences has been prepared. NL 1957 and NL 1961 are in English only, therefore references are to the English paragraphs alone in NL 1960.

2.7 Comparison of TS_x and NL 1979

TS_x is written on paper of 203mm x 329mm equivalent to half foolscap (8¼ x 13½ ins.). The typesize is 10 characters per inch which is also an Imperial standard. This would correspond with a copy made in the UK or USA, and would be compatible with the suggestion that Schwayder made the typescript in the USA in 1950 (cf. p.30).

NL 1979 used the Russell Version and the published text is the verbatim content of TS_x, including the first diagram. From this we may deduce that the content of TS_x is the content of TS 201a-x, the Russell Version. The reason for expressing it this way is that the physical characteristics of TS_x, when compared to descriptions of 201a-(1-3), still do not match.

There follows a line-by-line comparison of TS_x and NL 1979. Attention has been paid to all differences, except where underlining in TS_x has been rendered in italics in NL 1979, double underlining has been rendered in small capitals, and where non-English words have been italicised in NL 1979. TS_x contains a typewritten form of the Greek phi created by typing upper case O and over-typing /, for which the visually similar character Ø has been substitutes below. A full transcription of TS_x appears in volume 2.

TS_x (unpaginated f., typescript title page) reads: Notes on Logic by Ludwig Wittgenstein September 1913
NL 1979 headed: NOTES ON LOGIC by Ludwig Wittgenstein 1913

TS_x (unpaginated f.ii headed SUMMARY) ¶1 has no reference added
NL 1979 p.93, ¶1 ends: [Cf. 5.43]

TS_x (unpaginated f.ii headed SUMMARY) ¶3 has no reference added
NL 1979 p.93 ¶3 ends: [See 4.063]

TS_x (unpaginated f.ii headed SUMMARY) ¶6 has no reference added
NL 1979 p.93 ¶6 ends: [Cf. 2.0201]

TS_x (unpaginated f.ii headed SUMMARY) ¶7 footnote * reads: Russell - for instance...
NL 1979 p.93 ¶7 footnote 1 reads: Russell for instance...

TS_x (unpaginated f.ii headed SUMMARY) ¶8 has no reference added
NL 1979 p.93 ¶8 ends: [Cf. 3.315]

TS_x (unpaginated f.ii headed SUMMARY) ¶9 for both occurrences reads: Socrates and Plato...
NL 1979 p.93 ¶9 reads: "Socrates" and "Plato"

TS_x (unpaginated f.ii headed SUMMARY) ¶10 reads: the question "what is negated"...
NL 1979 p.94 ¶1 reads: the question what is negated...

TS_x (unpaginated f.ii headed SUMMARY) ¶11 reads: p implies "p is false" and includes the handwritten insertion of the sign ~
NL 1979 p.94 ¶2 reads: ~p implies "p is false"

TS_x (unpaginated f.ii headed SUMMARY) ¶12 footnote * reads: W-F = Wahr-Falsch.
NL 1979 p.94 ¶3 footnote 1 reads: W-F = Wahr-Falsch—i.e. True-False.

TS_x (unpaginated f.ii headed SUMMARY) ¶12 reads: Now what the Wf scheme does...
NL 1979 p.94 ¶3 reads: Now what the WF scheme does...

TS_x f.2 ¶3 typescript runs off the right-hand edge of the folio and reads: all those function
NL 1979 p.94 ¶6 reads: all those functions

TS_x f.2 ¶4 reads: (Note by B.R. ab means the same as WF, which means true-false.?)
NL 1979 p.94 omits note

TS_x f.2 footnote * reads: (It does not say p)... the old poles are correlated to p and includes two spaces for the handwritten insertion of the sign ~. These signs are omitted.
NL 1979 p.94 footnote 2 includes one insertion of the sign ~ and reads: (It does not say ~p)... the old poles are correlated to p.

TS_x f.2 ¶7 does not include a translator's comment
NL 1979 p.95 footnote 1 reads: I.e. sich verhalten, are related. Edd.

TS_x f.2 ¶7 does not include a translator's comment
NL 1979 p.95 footnote 2 reads: I.e. sich verhält, is related. Edd.

TS_x f.2 ¶9 does not include a reference
NL 1979 p.95 ¶4 ends: [Cf. 4.0621 and 5.5422]

TS_x f.2 ¶10 reads: The sense of an ab function of a proposition is a function of its sense
NL 1979 p.95 ¶5 reads: The sense of an ab function of a proposition is a function of its sense

TS_x f.3 ¶1 does not include a reference
NL 1979 p.95 ¶5 ends: [Cf. 5.2341]

TS_x f.3 ¶3 reads: Same objections in the case of apparent variables to the usual indefinables
NL 1979 p.96 ¶1 reads: Same objections in the case of apparent variables to old indefinables

TS_x f.3 ¶3 reads: The Notation is...
NL 1979 p.96 ¶2 reads: The notation is...

TS_x f.3 ¶3 indented logical notation reads: (x)Øx: a-(x)-aØxb-( x)-b and for ( x)x: a-( x)-aØxb-(x)-b and includes the handwritten insertion of the sign ∃ in three places.
NL 1979 p.96 ¶2 reads: (x)φx: a-(x)-aφxb-(∃x)-b and for (∃x)φx: a-(∃x)-aφxb-(x)-b

TS_x f.3 ¶5 reads: the fact that the symbol a stands in a certain relation to the symbol b...
NL 1979 p.96 ¶4 reads: the fact that the symbol "a" stands in a certain relation to the symbol "b"...

TS_x f.3 ¶5 does not include a reference
NL 1979 p.96 ¶4 ends: [Cf. 3.1432]

TS_x f.3 ¶7 does not include a reference
NL 1979 p.96 ¶6 ends: [Cf. 3.332]

TS_x f.3 ¶11 reads: (Components are forms and constituents.)
NL 1979 p.96 ¶10 reads: [Components are forms and constituents.]

TS_x f.3 ¶12 reads: Tale (Ø).Ø!x.
NL 1979 p.96 ¶11 reads: Take (φ).φ!x.

TS_x f.3 ¶12 reads: for which Ø! stands
NL 1979 p.96 ¶11 reads: for which "φ!" stands

First MS

TS_x f.4 ¶2 reads: (which are symbols having reference to facts)
NL 1979 p.97 ¶1 reads: [which are symbols having reference to facts]

TS_x f.4 ¶2 does not include a reference
NL 1979 p.97 ¶1 ends: [Cf. 2.141 and 3.14]

TS_x f.4 ¶3 does not include a reference
NL 1979 p.97 ¶2 ends: [Cf. 3.322]

TS_x f.4 ¶4 does not include a reference
NL 1979 p.97 ¶3 ends: [Cf. 3.143]

TS_x f.4 ¶5 reads: "( x,Ø).Øx" or "( ).xRy"... (e.g.) " ( x,y)xRy"? and includes the handwritten insertion of the signs ∃ ∃x,y ~ ∃
NL 1979 p.97 ¶4 reads: "(∃x,φ).φx" or "(∃x,y).xRy"... (e.g.) "~(∃x,y)xRy?

TS_x f.4 ¶6 reads: why " Socrates" means nothing is that " x" does not express... and includes two handwritten insertions of the sign ~
NL 1979 p.97 ¶5 reads: why "~Socrates" means nothing is that "~x" does not express...

TS_x f.4 ¶8 reads: Positive and negative facts there are, but not true and false facts
NL 1979 p.97 ¶7 reads: Positive and negative facts there are, but not true and false facts

TS_x f.4 ¶9 reads: (We might then say e.g. that "q" signifies in the true way what "not-q" signifies in the false way)
NL 1979 p.97 ¶8 reads: (We might then say e.g. that "q" signifies in the true way what "not-q" signifies in the false way)

TS_x f.4 ¶9 reads: if by "q" we mean "not-q"
NL 1979 p.97 ¶8 reads: if by "q" we mean "not-q"

TS_x f.4 ¶9 reads: we can mean the same by "q" as by "not-q"
NL 1979 p.97 ¶8 reads: we can mean the same by "q" as by "not-q"

TS_x f.4 ¶9 ends with the handwritten references: Cf. 4.061, 4.062, 4.0621
NL 1979 p.98 ¶1 ends: [Cf. 4.061, 4.062, 4.0621]

Second MS

TS_x f.5 ¶1 ends with the handwritten insertion: Cf. 4.02, 4.021, 4.027
NL 1979 p.98 ¶2 ends: [Cf. 4.02, 4.021, 4.027]

TS_x f.5 ¶3 reads: (particular)
NL 1979 p.98 ¶4 reads: [particular].

TS_x f.5 ¶5 ends with the handwritten reference: Cf. 4.024
NL 1979 p.98 ¶6 omits reference.

TS_x f.5 ¶6 ends with the handwritten reference: Cf. 4.024
NL 1979 p.98 ¶7 reads: [Cf. 4.024]

TS_x f.5 ¶7 ends with the handwritten reference: Cf. 3.144
NL 1979 p.98 ¶8 omits reference.

TS_x f.5 ¶9 reads: in " p", " " looks like "Ø" but is not like it. and includes two handwritten insertions of the sign ~
NL 1979 p.98 ¶10 reads: in "~p", "~" looks like "φ" but is not like it.

TS_x f.5 ¶9 ends with the handwritten reference: Cf. 4.024
NL 1979 p.98 ¶10 reads: [Cf. 4.024]

TS_x f.6 ¶2 does not include a reference
NL 1979 p.99 ¶6 ends: [Cf. 4.024]

TS_x f.6 ¶3 reads: (Thus e.g. no proposition is indefinable)
NL 1979 p.99 ¶7 reads: [Thus e.g. no proposition is indefinable]

TS_x f.6 ¶4 includes text deleted and bracketed by hand and reads: all propositions that have sense [? meaning]
NL 1979 p.99 ¶8 reads: all propositions that have sense

Third MS

NL 1979 p.99 footnote 1 has no correspondence in TS_x

TS_x f.7 ¶1 reference Cf. 4.063 inserted by hand
NL 1979 p.100 ¶1 reads: [Cf. 5.132].

TS_x f.7 ¶3 ends with the handwritten reference: Cf. 5.132
NL 1979 p.100 ¶3 omits reference.

TS_x f.7 ¶5 reads: (Urzeichen)
NL 1979 p.100 ¶5 reads: (Urzeichen)

TS_x f.7 ¶5 the word "questioning" deleted and marginal insertion by hand which reads: proportion
NL 1979 p.100 ¶5 reads: questioning

TS_x f.7 ¶5 reads: p q is this: p.p q. q and includes three handwritten insertions of the sign ⊃ and one of the sign _p.q
NL 1979 p.100 ¶5 reads: p⊃q is this: p.p⊃q.⊃_p.q.q

TS_x f.7 ¶6 reads: everything else
NL 1979 p.100 ¶6 reads: everything else

TS_x f.7 ¶6 ends with the handwritten reference: Cf. 5.43
NL 1979 p.100 ¶6 reads: [Cf. 5.43]

TS_x f.8 ¶2 ends with the handwritten reference: Cf. 2.0201
NL 1979 p.101 ¶2 reads: [Cf. 2.0201]

TS_x f.8 ¶3 ends with the handwritten reference: See 5.42
NL 1979 p.101 ¶3 reads: [Cf. 5.42]

TS_x f.8 ¶4 reads: p ( x).Ø(x)=p and includes the handwritten insertion of the signs ^ [ ∃ ]
NL 1979 p.101 ¶4 reads:p̂{(∃x).φ(x)=p}

TS_x f.8 ¶5 ends with the handwritten reference: Cf. 3.315
NL 1979 p.101 ¶5 reads: [Cf. 3.315]

TS_x f.8 ¶6 ends with the handwritten reference: Cf. 4.1241
NL 1979 p.101 ¶6 reads: [Cf. 4.1241]

TS_x f.8 ¶9 ends with the handwritten reference: See 3.144
NL 1979 p.102 ¶1 reads: [Cf. 3.144].

TS_x f.8 ¶10 omitted word "by" inserted by hand
NL 1979 p.102 ¶2 reads: the form of propositions aRb by saying...

TS_x f.8 ¶11 mistyping which reads: "p2 over-typed to read: "p"

TS_x f.8 ¶11 reads: "b p" and includes the handwritten letter a in superscript
NL 1979 p.102 ¶3 reads: "bͣp"

TS_x f.9 ¶3 ends with the handwritten reference: See 3.203
NL 1979 p.102 ¶5 reads: [Cf. 3.203].

TS_x f.9 ¶6 reads: p q and includes the handwritten insertion of the signs ~ ∨ ~
NL 1979 p.102 ¶8 reads: ~p∨~q

TS_x f.9 ¶7 reads: the definition of " " by " " and " ", or of " " by " " and " "? and includes the handwritten insertion of the signs ⊃ ∨ ~ ∨ ~ ⊃
NL 1979 p.103 ¶1 reads: the definition of "⊃" by "∨" and ".", or of "∨" by "." and "⊃"?

TS_x f.9 ¶7 reads: "p/q (i.e. p )" and includes handwritten insertion of the signs ~ ∨ ~q
NL 1979 p.103 ¶1 reads: "p│q (i.e. ~p∨~q)"

TS_x f.9 ¶7 reads: "p/q"
NL 1979 p.103 ¶1 reads: "p│q"

TS_x f.9 ¶7 reads " " and includes handwritten insertion of the sign │
NL 1979 p.103 ¶1 reads: "p│q"

TS_x f.9 ¶7 reads: "p q" and includes handwritten insertion of the sign │
NL 1979 p.103 ¶1 reads: "p│q"

TS_x f.9 ¶7 reads " " and includes handwritten insertion of the sign │
NL 1979 p.103 ¶1 reads: "p│q"

TS_x f.9 ¶7 reads "p q" and includes handwritten insertion of the sign │
NL 1979 p.103 ¶1 reads: "p│q"

TS_x f.9 ¶7 ends with the handwritten reference: See 5.44
NL 1979 p.103 ¶1 reads: [Cf. 5.44].

TS_x f.9 ¶8 reads " " and includes handwritten insertion of the sign _
NL 1979 p.103 ¶2 reads: "_"

TS_x f.9 ¶8 ends with the handwritten reference: See 4.42
NL 1979 p.103 ¶2 reads: [Cf. 4.442].

TS_x f.9 ¶9 ends with the handwritten reference: 5.5422
NL 1979 p.103 ¶3 reads: [See 5.5422].

TS_x f.10 ¶1 ends with the handwritten reference: Cf. 4.024
NL 1979 p.103 ¶4 reads: [Cf. 4.024].

TS_x f.10 ¶2 reads: "p q", "p q", etc. are only then not provisional when " " and " " stand within the scope of a generality-sign (apparent variable) and includes handwritten insertion of the signs ∨ ⊃ ∨ ⊃
NL 1979 p.103 ¶5 reads: "p∨q", "p⊃q", etc. are only then not provisional when "∨" and "⊃" stand within the scope of a generality-sign [apparent variable]

Fourth MS

TS_x f.11 ¶1 includes typescript remark: I doubt this. R? DS.¹⁷
NL 1979 p.103 ¶6 omits remark

TS_x f.11 ¶1 ends with the handwritten reference: See 4.26
NL 1979 p.103 ¶6 reads: [Cf. 4.26].

TS_x f.11 ¶2 ends with the handwritten reference: Cf. 4.0621
NL 1979 p.103 ¶7 reads: [Cf. 4.0621].

TS_x f.11 ¶3 reads: "Øa a=a" and includes handwritten insertion of the signs ⊃(φ,a).
NL 1979 p.104 ¶1 reads: "φa. ⊃_φ,aa=a."

TS_x f.11 ¶4 handwritten correction of "created" to "creates"
NL 1979 p.104 ¶2 reads: creates

TS_x f.11 ¶4 ends with the handwritten reference: Cf. 5.123
NL 1979 p.104 ¶2 reads: [Cf. 5.123].

TS_x f.11 ¶5 reads: "p p" and includes handwritten insertion of the signs ∨ ~
NL 1979 p.104 ¶3 reads: "p∨~p"

TS_x f.11 ¶5 reads: "(p). p p" and includes handwritten insertion of the signs ∨ ~
NL 1979 p.104 ¶3 reads: "(p).p ∨ ~p"

TS_x f.11 ¶5 ends with the handwritten reference: Cf. 4.461
NL 1979 p.104 ¶3 reads: [Cf. 4.461].

TS_x f.11 ¶6 reads: all x"s and y's
NL 1979 p.104 ¶4 reads: all x's and y's

TS_x f.11 ¶6 ends with the handwritten reference: Cf. 4.024
NL 1979 p.104 ¶4 reads: [Cf. 4.024].

TS_x f.11 ¶8 reads: in regard to
NL 1979 p.104 footnote 1 comments on this translation

TS_x f.11 ¶8 reads: (the facts)... ("gleichsinnig")... (entgegengesetzt")
NL 1979 p.104 ¶6 reads: [the facts]... ["gleichsinnig"]... [entgegengesetzt"]

TS_x f.11 ¶9 reads: ’"p" is true’ p;... " " and includes two handwritten insertion of the sign ≡
NL 1979 p.104 ¶6 reads: "’p’ is true"≡p;... "≡"

TS_x f.11 ¶10 ends with the handwritten reference: See 5.2341
NL 1979 p.105 ¶2 reads: [Cf. 5.2341].

TS_x f.11 ¶12 ends with the handwritten reference: See 3.142
NL 1979 p.105 ¶4 omits reference.

TS_x f.12 ¶2 originally read: ...true or false. amended by hand to read: ...true or false, not both.
NL 1979 p.105 ¶6 reads: ...true or false, not both.

TS_x f.12 ¶3 reads: This was typed in but had exesses through it. (D.S.)¹⁸
NL 1979 p.105 omits this remark.

TS_x f.12 ¶4 ends with the handwritten reference: Cf. 5.1241
NL 1979 p.105 ¶7 reads: [Cf. 5.1241].

TS_x f.12 ¶5 reads: ( x. y).xRy and similar ones... ( x,y).Ø(x.y) and includes two handwritten insertion of the sign ∃ and two corrections of "." to ","
NL 1979 p.105 ¶8 reads: (∃.x.y).xRy and similar ones... (∃x,y).φ(x,y). If the first correction had been observed the first expression would have read: (∃.x,y).xRy

TS_x f.12 ¶5 originally read: But when we introduce... amended by hand to read: But when we introduced
NL 1979 p.105 ¶8 reads: But when we introduce...

TS_x f.12 ¶6 reads: ( x,y).xRy and others... and includes the handwritten insertion of the sign ∃
NL 1979 p.105 ¶9 reads: (∃x,y).xRy and others...

TS_x f.12 ¶6 ends with the handwritten reference: Cf. 5.451
NL 1979 p.105 ¶8 reads: [Cf. 5.451].

TS_x f.12 ¶9 reads: if we say "A believes that ’p’ is true"
NL 1979 p.106 ¶3 reads: if we say "A believes that ’p’ is true".

TS_x f.12 ¶9 reads: "A believes ’p’"
NL 1979 p.106 ¶3 reads: "A believes ’p’".

TS_x f.12, the diagram occurs after ¶9 which ends "...a-p-b." and before ¶10 which begins "The epistemological questions...".
NL 1979 p.106 publishes the diagram after ¶3, the last sentence of which begins "The epistemological questions..."

TS_x f.13 ¶1 ends with the handwritten reference: 3.143
NL 1979 p.106 ¶5 reads: [Cf. 3.1432].

TS_x f.13 ¶4 ends with the handwritten reference: Cf. 4.111
NL 1979 p.106 ¶8 omits reference.

TS_x f.13 ¶6 ends with the handwritten reference: See 4.1121
NL 1979 p.106 ¶10 reads: [Cf.4.1121].

TS_x f.13 ¶8 handwritten reference inserted after "...different sense" which reads: 4.032 and continues in typescript which reads: Crossed out but originally typed in (D.S.).
NL 1979 p.106 ¶12 reads: [Cf.4.032].

TS_x f.13 ¶10 ends with the handwritten reference: Cf. 4.113
NL 1979 p.106 ¶14 omits reference.

TS_x f.13 ¶11 ends with the handwritten reference: See 4.111
NL 1979 p.106 ¶15 reads: [Cf.4.111].

TS_x f.13 ¶13 ends with the handwritten reference: Cf. 4.002
NL 1979 p.107 ¶2 omits reference.

TS_x f.13 ¶14 ends with the handwritten reference: See 3.144
NL 1979 p.107 ¶3 omits reference.

TS_x f.13 ¶15 reads: (Urbilder von Zeichen)
NL 1979 p.107 ¶4 reads: [Urbilder von Zeichen].

TS_x f.13 ¶16 has a marginal handwritten note which reads: wrong termed? This probably refers to the wording of the last sentence which reads: ...Russell's theory of manufactured relations
NL 1979 p.107 ¶5 ignores the handwritten remark and reads: ...Russell's theory of manufactured relations

TS_x f.14 ¶1 ends with the handwritten reference: 6.12
NL 1979 p.107 ¶8 omits reference.

TS_x f.14 ¶2 ends with the handwritten reference: Cf. 3.332
NL 1979 p.107 ¶9 reads: [Cf.3.332].