The discovery of E.W.

Beth’s semantics for

intuitionistic logic
A.S. Troelstra and P. van Ulsen

Abstract
One of van Benthem’s predecessors, and the first in line of the Dutch logicians, was
Evert Willem Beth. During the last decade of his life he worked on a truly constructive
semantics for intuitionistic logic, with a corresponding completeness theorem. The
result is known as “Beth models”. We try to describe his intents and efforts, but
it is not possible to give a clear story line: the data are too scarce. However, we
attempt a reconstruction. For this we not only used published records, but as much
as possible also quotes from correspondence. Beth semantics combine trees, tableaus,
choice sequences and fans. Intuitionistically acceptable completeness required Beth
to avoid certain classical notions, in particular König’s Lemma. Instead, Beth used
Brouwer’s fan theorem.

Contents
1 Introduction 2

2 Semantics for intuitionistic logic “before Beth” 2

3 Beth’s Indagationes paper from 1947 and the

Berkeley lecture in 1952 3

4 On the Paris lecture of 1955 and the tools for constructive

semantics 4

5 The paper from 1956 and afterwards 7

1
1 Introduction
Johan van Benthem occupies a position which was once held by Evert Willem
Beth. Johan’s 50th birthday seemed a suitable occasion for investigating the
work of one of his scientific ancestors; so here we take a look at Beth’s discovery
of “Beth semantics” and completeness for “Beth models”.
To our regret, we do not have a clear story line: the data are too scarce. But
we have done our best, and have sketched a reconstruction. For this reconstruc-
tion, we used not only published records, but as much as possible also quotes
from correspondence. The corrrespondence is preserved at Haarlem (Rijksar-
chief in Noord-Holland) with the name of “E.W. Beth Archive”.

2 Semantics for intuitionistic logic “before Beth”

In L.E.J. Brouwer’s writings from 1908 onwards there is, more or less implicitly,
an informal intuitionistic interpretation of certain logical operators such as “or”,
“implies”, and “not”.
This interpretation was formulated more precisely, as an informal interpre-
tation of formal logical expressions by A. Heyting, first in Heyting (1931), and
later, more completely in the monograph Heyting (1934). A similar interpre-
tation was formulated by A. Kolmogorov in a paper which appeared in 1932.
Originally Heyting and Kolmogorov regarded their respective interpretations
as distinct, but later Heyting came to regard them as essentially the same. It
is for this reason that this interpretation is usually referred to as the Brouwer-
Heyting-Kolmogorov interpretation, or BHK-interpretation.
Where classical semantics describes how the truth-value of a logically com-
pound statement is determined by the truth-values of its components, the BHK-
interpretation describes what it means to prove a logically compound state-
ment in terms of what it means to prove the components. In this explanation,
“constructive proof”, “constructive method” appear as primitive notions. Kol-
mogorov interpreted statements as problems, and the solving of the problem
associated with a logically compound statement is explained in terms of what
it means to solve the problems represented by the components. The connection
with Heyting’s formulation becomes clear if we think of “solving the problem
associated with statement A” as “proving A”.
Because of the vagueness of the notions of “constructive proof”, “construc-
tive operation”, the BHK-interpretation has never become a versatile technical
tool in the way classical semantics has. Perhaps it is correct to say that by
most people the BHK-interpretation has never been seen as an intuitionistic
counterpart to classical semantics.
Another approach to semantics for intuitionistic logic started with Jaskowski
(1936) who used certain systems of truth-values, special cases of Heyting alge-
bras, for a completeness proof for intuitionistic propositional logic.
Stone (1937) does not explicitly state a completeness theorem, but in fact
his paper contains a completeness result for intuitionistic propositional logic.
Tarski (1938) obtains a completeness result for intuitionistic propositional cal-

2
culus IPC w.r.t. valuations using open sets in topological spaces (in particular
IRn ) as values. Tarski also sketches an extension to valuations in a suitable
class of lattices (More about this in Troelstra & van Dalen (1988), p. 731). The
Kripke-tree semantics Beth turned to in the final years of his life belongs in this
category too. It seems to have no direct relation to the meaningful semantics
discussed in the present article (see de Jongh & van Ulsen (1998)).
Finally, in 1945 S.C. Kleene published an interpretation (realizability) of
formalized intuitionistic arithmetic in the structure of the natural numbers.
This was truly a semantics, linking formal expressions to truth in a familiar
structure.
It is against this background that Beth’s interest in an intuitionistically
meaningful semantics for intuitionistic predicate logic arose. But Beth was not
himself an intuitionist as it appears in a letter to K. Schröter of July 16, 1960:
In your letter you call me a “follower of the intuitionistic foundations of
mathematics”. Although I am strongly interested in intuitionism, and know
this point of view for geographical reasons very well, more than other work-
ers in the field, I do not want to be considered as a follower, let alone as an
exponent. I would describe my point of view as tolerant-classical.

3 Beth’s Indagationes paper from 1947 and the

Berkeley lecture in 1952
According to Beth’s own statement in Beth (1959b), p. 15, his concern with the
semantics of intuitionistic logic dates from 1943. In 1946 Beth was appointed in
Amsterdam as the first Dutch professor in logic. Already in 1947 Beth published
a paper in the Proceedings of the Royal Dutch Academy of Sciences with the
title “Semantic considerations on intuitionistic mathematics.”
The discussion in this paper is of a very general nature; Beth explains at
some length the idea of semantics in Tarski’s sense. He then turns to intu-
itionism; as its key notion he sees the notion of a spread. He cites Brouwer’s
definition and then recasts it in a more formal way. Later Heyting, in his
book “Intuitionism, an introduction” gave a simplified and clear presentation
of Brouwer’s notion. Beth’s reformulation may be seen as a first step in this
direction. (This point was appreciated by S.C. Kleene in his review of Beth’s
paper in the Journal of Symbolic Logic in 1948.) Beth also asks whether the
spread law could always to be taken to be recursive; he thought not, from an
intuitionistic point of view. Recursiveness was to narrow for intuitionistic pur-
poses. Beth got his credits, together with Kleene, in Odifreddi (1989), p. 118.
But on the whole the content of this paper is quite meager, although it should
not be forgotten that Tarski’s notion of semantics dating from 1935 was at that
time still relatively unfamiliar to many.
There exists a draft in Dutch of this paper, which apparently was sent
together with a letter to L.E.J. Brouwer; the letter is dated july 7, 1945. Cu-
riously enough, the letter starts with “Herewith I take leave to send you some
remarks on non-intuitionistic mathematics [ . . . ]”, while the manuscript which
is supposed to have been sent with this letter almost exclusively deals with

3
intuitionism. So is this a typing error, or do the manuscript and the letter not
really belong together? Perhaps neither of these: it probably was a joke, typical
for Beth’s sense of humor, according Dick de Jongh.
Beth returned to spreads, and in particular Brouwer’s “fan theorem” for
finitary spreads in the lecture “Compactness proofs in intuitionistic mathemat-
ics” at the Mathematics Colloquium at Berkeley (May 15, 1952; ms. in two
versions). In this lecture Beth presents a version of Brouwer’s proof of the bar
theorem and the fan theorem; Beth states that the version is due to a student
of Heyting, Pinxter. In the argument as presented by Beth, some subtleties of
Brouwer’s original argument have been swept under the carpet. Beth then pro-
ceeds to show that the fan theorem can fulfill the role of classical compactness
in intuitonistic mathematics; he discusses in particular how the Heine-Borel
covering theorem for the interval may be obtained from the fan theorem. He
realizes that the application of the fan theorem is not entirely straightforward,
but that one must choose the right sort of spread representation for the inter-
val (a fact known to Brouwer). In short, the paper does not contain any new
result, but it testifies to Beth’s ongoing preoccupation with Brouwer’s proof of
the fan theorem and shows his awareness of the fact that if one wants to find
an intuitionistic substitute for a classical compactness argument, one needs the
fan theorem. It is remarkable that Beth never — not in this lecture but also
not in his later work — referred to Brouwer (1926). There Brouwer presented
an intuitionistically acceptable variant of Heine-Borel.

4 On the Paris lecture of 1955 and the tools for con-

structive semantics
Beth made real progress with the semantics of intuitionistic predicate logic only
after he had invented the semantic tableaus for classical logic in 1954 (later pub-
lished as Beth (1956a)). In Beth (1955a), the next step, he tried to formulate a
decision method for intuitionistically acceptable formulas, given that they are
classically acceptable. He used semantic tableaux, and their deductive tran-
formation. Beth (1955b), p. 341–342: “I stated the conjecture that semantic
tableaux might provide us a decision procedure for intuitionistic validity of clas-
sically valid formulas.” Beth’s hope was destroyed by Kleene’s letter of June 7,
1955 (Madison). Kleene gave some counterexamples and finished his objections
as follows:
There can be no decision procedure for the problem ‘let A be a formula
which holds classically; does A also hold intuitionistically?’ [ . . . ] At any
rate I do not see how to rule out such possibilities [i.e. Kleene’s counterex-
amples]; and so it does not seem implausible that increasing the number
of individuals taken into consideration [as Beth proposed] might give room
for a correct proof intuitionistically in place of one correct classically but
incorrect intuitionistically.
For Beth’s agreement with Kleene see Beth (1955b), p. 341–342.
The intuitionistic system G3 in Kleene’s “Introduction to Metamathemat-
ics” (p. 481), was taken by Beth as the basis for his syntax. Out of this Beth

4
developed his tableaux (trees with attached formulas); or, as Beth said in his
“Foundations of mathematics”, p. 450,
If read upside down, the above [i.e. intuitionistic syntactical] rules may also
be construed as instructions for the construction of a semantic tableau.
Kleene, and his predecessor G. Gentzen, had disjunctive and conjunctive clauses
(for the tableaux the conjunctive and disjunctieve splittings). Beth added in-
finite regression for the branches of his trees; this was produced by cyclic per-
mutation of formulas.
It is not very clear how Beth discovered his semantics. Correspondence
about this topic is of a later date and Beth never gave a precise description of
the way he arrived at his ideas. So we use a lot of quotations from sources later
than 1956, but a real reconstruction is impossible.
On May 3, 1957 Beth sent a letter to Heyting from Baltimore, while he was
a visiting professor at Johns Hopkins University. In that letter, written after
his discovery, you can find a clue to the beginnings of the intuitionistic use of
semantic tableaux:
Indeed I start from a classical, hence intuitionistically debatable notion of
model; however, my considerations concern classical logic. Moreover, a con-
siderable part of the construction is nevertheless intuitionistically accept-
able, since the tree construction associated with a logical problem produces
a collection of models which can be represented by a finitary spread.
In the same letter Beth told that he was examined by G. Kreisel:
Concerning my construction of intuitionistic logic, I had in Princeton an in-
teresting conversation with Kreisel, who seemed rather taken with the idea;
he proposed an exam-question, which I was able to answer satisfactorily.
In 1955 Beth gave a lecture in Paris on the semantics of intuitionistic logic,
in which he sketches the essentials of his semantics. The lecture was published
in 1958, only after his more substantial paper on the same topic had appeared
in 1956.
The lecture of 1955 is a big step forward. Beth sketches the clauses for intu-
itionistic validity in his models, and discusses some examples. However, there
is still no indication of a completeness proof, and one of the clauses for validity
is manifestly inadequate (namely, the clause for the validity of an existential
statement). The absence of a completeness proof is not so strange, if you read
Beth’s letter of September 5, 1955 to A. Robinson (italics by the authors):
In my paper on intuitionistic logic, I take a non-intuitionistic attitude, and
convince myself that, if an intuitionist uses a law of logic not contained
in Heyting’s system, then I can find an intuitionistic counter-example to
prove that this law is not acceptable from an intuitionist point of view. For
the denumerable case, the lemma of infinity is not needed, we only need
the (metamathematical) principle of the excluded third. As for the classi-
cal logic, an entirely constructive completeness proof cannot be given. But
the counter-example, the existence of which is proved by non-constructive
methods, is in itself constructive.
Within a year Beth changed his mind. His next step, in 1956, was the develop-
ment of semantics with constructive completeness for intuitionistic logic. For

5
his formulation of validity Beth made use of trees because of their relationship
with tableaux. In Beth (1956b), p. 379, he motivates his use of choice sequences
as follows (italics by Beth):
[ . . . ] the introduction of the notion of a choice sequence brings a subjective
element into the situation. But, as Heyting rightly observes, this subjective
element is eliminated if we agree to concentrate upon such properties of
choice sequences as appear after a finite number of choices. This attitude
implies, however, a radical change in the semantic notions. The classical
rules determine [ . . . ] the validity or non-validity of a formula on each
branch separately [ . . . ] These difficulties vanish, if we agree to determine
validity or non-validity, not on individual branches, but collectively on all
those branches which have a certain initial element in common, that is, on
a subtree.
In this passage, he explains his use of “submodel” (bar).
Beth had to avoid some classical principles as compactness (ms. E.W. Beth,
“Semantische Tafeln für die intuitionistische Prädikatenlogik erster Ordnung”,
lecture IV. Österr. Mathematikerkongress, Wien, September 17–22, 1956):
If the construction of tableaux is interpreted as an attempt to construct
a counter-model, then just as before we have to use an intuitionistic non-
acceptable compactness argument for the proof that for every non-closed
semantic tableau a corresponding counter-model actually exists.
The original semantic tableaux had been countermodel-constructions. For intu-
itionistic purposes these were now transformed into an operation of ‘fitting in’,
as witnessed by a letter from Beth to Heyting, dated November 18, 1955 (this
letter must have been written after his Paris lectures, but nevertheless he did
not modify Beth (1958) in this respect):
Let us think of a certain (given) model of the A’s, and also of a model, as
yet undetermined, which is a union of submodels as intended above. Call
these models M and N . Now we are going to split M and N ever farther
into submodels. Splitting of M yields a conjunctive splitting of the tableau,
splitting of N a disjunctive splitting. The construction succeeds, if we have
split M into parts, each of which fits into some piece of N .
The two “truth values” here are “true” and “not yet decided”. The combination
< M, N > is a fan. Beth used Brouwer’s fan-theorem with its restrictions as a
framework for his intuitionistic completeness proof.
It is possible to have infinite branches in a submodel; these are produced
by Beth’s cyclic permutation of formulas. Beth introduced trunks (truncated
trees) with a finite depth, in order to compare the conjuctive and the disjunctive
part of a model up to a certain depth.
Beth formulated it in his lecture “Intuitionistic predicate logic” at Cornell
University in 1957 as follows:
This [effective] procedure [from M to N as above] cannot applied to com-
plete trees, it must work on trunks of sufficient length. As trunks of models
cannot (always) be distinguished from trunks of semi-models, the procedure
must work as well on trunks of semi-models which are sufficiently like mod-
els. [ . . . ] the procedure must lead to one of the following results: (1) the
trunk cannot be extended into a model of K [i.e., the conjunctive part]; (2)
every extension of the trunk into a model of K is also a model of L [i.e., the

6
 disjunctive part]. hence, duplicating Brouwer’s second proof of the so-called
fundamental theorem of finitary spreads, we can establish the existence of
an upper bound k 0 for all “sufficient lenghts”.

5 The paper from 1956 and afterwards

Beth published in 1956 a paper (some parts of which we used already) in the
Proceedings of the Royal Dutch Academy of Sciences in which he gives a correct
definition of validity for his models, and presents a completeness proof in two
versions, a classical and an intuitionistic one. Beth’s arguments are not very
easy to follow; in part this is due to a somewhat heavy notation. His classical
proof was accepted as in essence correct. But several critics complained about
the purported intuitionistical argument. They either considered it as unintelli-
gible (Heyting) or flawed. Improving Beth’s proposals took a lot of time, but
as Kreisel said to Beth in his letter of December 19, 1957 (Reading):
I regard your work as a very important contribution to the subject. For this
reason it seems to me desirable to give a flawless proof of your result [i.e.
‘duplicating Brouwer’s proof’].
After nearly three years Kreisel told Beth in a letter from Paris, dated November
11, 1960, about his work:
During the summer we prepared at Stanford a Project report concerning the
completeness of intuitionistic logic including (a) [ . . . ], (c) a very detailed
version of your completeness proof with a full proof of its applicability to
decidable subsystems, (d) [ . . . ]
Under (c) one finds the reference to Dyson & Kreisel (1961). Beth’s reaction
was a short, laconic one (letter to Kreisel, December 5, 1960):
I shall be very much interested to see your Project report. If I do not find
time to read it carefully, it would certainly provide excellent material for a
seminar either by myself or by Heyting and me.

We have not been able to find any later remark by Beth on this report.
Beth never took the time to thrash out the criticisms in detail. In the
beginning he claims that the criticism of his proof should really be taken as
criticism of Brouwer’s argument for the fan theorem. An example you can find
in Beth’s letter to G. Kreisel of December 23, 1957:
With respect to the intuitionistic proof, my own attitude is ambivalent.
This proof seems to belong to intuitionistic higher-order logic, for which no
formal standards are available at present. On the other hand, all proofs of
the fan theorem which I have seen seem to be fallacious. Looking at the
matter under this aspect, ‘duplicating Brouwer’s proof’ [a remark made by
Kreisel in his letter to Beth, (Reading), December 18, 1957 (p. 3)] is a polite
expression for imitating Brouwer’s fallacy.
Kleene and Kreisel had discussed Beth’s results. Heyting wrote in a letter to
Beth of January 19, 1958 (Notre Dame) as a significant remark:
It appeared that your paper on intuitionistic logic had led to a rather thor-
ough dicussion between Kleene and Kreisel. New to me in this was, that
your interpretation of negation and implication only holds if one permits

7
 absolutely free choice sequences, and not the sequences determined by a
law.
If the flaw in Beth’s alleged intuitionistic completeness proof did not reside
in the fan theorem per se, what was then the real weakness? Beth’s earlier
version (e.g., Beth (1959a), p. 459) of his intuitionistic completeness was: “If a
sequent C ⇒ D holds true intuitionistically, then it is derivable in the formal
system F 0 [i.e. Beth’s syntactical part].” At the end of the intuitionistic section
in his ‘Foundations’, p. 461, he changed this into: “If all models M which fulfil
the conjunctive C also fulfil the disjunctive D [of the sequent C ⇒ D], then
it [i.e. the sequent] is derivable in the formal system F 0.” The discussion
thereafter concerns the status of “all models”. Beth wrote December 8, 1957 to
Kreisel (at that time Beth thought his ‘Foundations’ was ready for printing):
The hypothesis in the completeness theorem contains the quantifier “all
models”, whereas (in general) the species of all models of a given conjunc-
tive cannot be represented by a spread. My paper contains a method for
embedding all models of a given conjunctive in a spread which in addition
contains semi-models.
But presumably the species of all models can also be embedded in other
spreads and, as I have shown in my paper in Amsterdam [i.e. Beth (1959b)]
the meaning of the quantifier “all models” may depend on the choice of the
spread in which these models are embedded.[. . . ] so the matter can only
be settled on the basis of a discussion on intuitionistic higher-order logic,
which seems to be an immensely intricate subject.
Not long after the first interchanges with G. Kreisel (and S.C. Kleene?) Beth
had come to realize that the crux of the matter was his application of the fan
theorem to a spread of initial parts of potential models; not all of these initial
parts belonged to proper models, some of them could only lead to inconsistent
models. In other words, “validity in all Beth-models” had to be read as “validity
in all potential models” in order to make sense of Beth’s argument.
Beth never found time to integrate critical remarks (Beth (1959a), p. 461:
“[. . . ] objections which have been raised from different viewpoints, by A. Heyt-
ing and by K. Gödel and G. Kreisel”) in his work. Certainly he did have no time
just before the printing of his ‘Foundations’; so he wrote to Kreisel, December
23, 1957 in relation to his “Foundations of mathematics”:
So it seems better to include a summary of my material in its present state
and to mention the fact that certain objections have been raised and my own
attitude on the above lines. I plan to go into the problem of intuitionistic
higher-order logic when the book is ready.
Beth never carried out these plans; so it was left to Veldman (1976) and de Swart
(1976), much later, to rediscover this idea and make use of it.

References
Beth, E.W. (1947), ‘Semantic considerations on intuitionistic mathematics’,
Indagationes Mathematicae 9, 572–577.
Beth, E.W. (1955a), ‘Remarks on natural deduction’, Indagationes Mathemat-
icae 17, 322–325. (dedicated to Robert Feys).

8
Beth, E.W. (1955b), ‘Semantic entailment and formal derivability’, Med. Konkl.
Nederl. Akad. v. Wetensch., Nieuwe Reeks 18(13), 309–342.
Beth, E.W. (1956a), l’Existence en mathématiques, number 10 in ‘Collection
de logique mathématiques, série A’, Gauthier-Villars/Nauwelaerts,
Paris/Louvain. (66 p.). (Beth’s lectures at Sorbonne University
(Paris), March 29 – April 2, 1954).
Beth, E.W. (1956b), ‘Semantic construction of intuitionistic logic’, Koninklijke
Nederlandse Akademie van Wetenschappen, Mededelingen, Nieuwe
Reeks 19(11), 357–388.
Beth, E.W. (1957), Intuitionistic predicate logic, in ‘Summaries of talks pre-
sented at the Summer Institute for Symbolic Logic in 1957’, Vol. 1,
Cornell University, pp. 113–121. (19602 ).
Beth, E.W. (1958), Construction sémantique de la logique intuitioniste, in
‘Le raisonnement en mathématique et en sciences experimentales
(Paris, 26 Sept.–1 Oct. 1955)’, number 70 in ‘Colloques Intern.
du C.N.R.S.’, pp. 77–83. ( A. Heyting, ‘Interventions’, related to
Beth’s lecture: p. 84).
Beth, E.W. (1959a), The foundations of mathematics. A study in the philosophy
of sciences, Studies in Logic, North Holland, Amsterdam. (XXVI
+ 722 pp.). (19652 revised, North Holland, (XXVIII + 741 pp.).
(19663 , Harper).
Beth, E.W. (1959b), Remarks on intuitionistic logic, in A. Heyting, ed., ‘Con-
structivity in mathematics’, Studies in Logic, (Colloquium con-
structivity in mathematics, Amsterdam, August 1957), North Hol-
land, Amsterdam, pp. 15–25.
Brouwer, L.E.J. (1926), ‘Die intuitionistische Form des Heine-Borelschen
Theorems’, Koninkl.Akad.Wetensch., Proceedings 29(6), 866–867.
(=De intuı̈tionistische vorm van het theorema van Heine-Borel,
K.A.W., Verslagen 35, pp. 667–678).
de Jongh, D.H.J. & van Ulsen, P. (1998), Beth’s nonclassical valuations, Tech-
nical Report LP-1998-12, Institute for logic, language and compu-
tation, Universiteit van Amsterdam.
de Swart, H.C.M. (1976), ‘Another intuitionistic completeness proof’, The Jour-
nal of Symbolic Logic 41, 644–662. (received September 24, 1974).
Dyson, V.H. & Kreisel, G. (1961), Analysis of Beth’s semantic construction of
intuitionistic logic, Technical Report 3, Applied mathematics and
statististical laboratories, Stanford University, Stanford. (Office of
Ordnance Research, Contract No. DA-04-200-ORD-997).
Heyting, A. (1931), ‘Die intuitionistische Grundlegung der Mathematik’, Erken-
ntnis 2, 106–115. (=Annalen der Philosophie 10).
Heyting, A. (1934), Mathematische Grundlagenforschung. Intuitionismus, Be-
weistheorie, Ergebnisse der Mathematik und ihre Grenzgebiete,
Springer, Berlin.

9
Heyting, A. (1956), Intuitionism, An introduction, Studies in logic, North Hol-
land. (19662 , 19713 , both revised).
Jaskowski, S. (1936), Recherches sur le système de logique intuitioniste, in
‘Actes du Congrès International de Philosophie Scientifique (Sor-
bonne, Paris, 1935), VI, Philosophie des Mathématiques’, Her-
mann, Paris, pp. 58–61.
Kleene, S.C. (1945), ‘On the interpretation of intuitionistic number theory’,
The Journal of Symbolic Logic 10, 109–124.
Kleene, S.C. (1948), ‘Review of E.W. Beth, ‘Semantical considerations on intu-
itionistic mathematics’, Indag. Math 9, 572–577’, Journal of Sym-
bolic Logic 13, 173.
Kleene, S.C. (1952), Introduction to metamathematics, number 1 in ‘Bibliotheca
Mathematica’, North Holland, Amsterdam. (19592 ).
Kolmogorov, A.N. (1932), ‘Zur Deutung der intuitionistischen Logik’, Mathe-
matische Zeitschrift 35, 58–65.
Odifreddi, P. (1989), Classical recursion theory, The theory of functions and
sets of natural numbers, number 125 in ‘Studies in Logic’, North
Holland, Amsterdam. (19922 ).
Stone, M.H. (1937), ‘Topological representations of distributive lattices and
Brouwerian logics’, Časopis Pro Pěstvovánı́ Matematiky a Fysiki
67, 1–25.
Tarski, A. (1935), ‘Der Wahrheitsbegriff in den formalisierten Sprachen’, Studia
Philosophica 1, 261–405.
Tarski, A. (1938), ‘Der Aussagenkalkül und die Topologie’, Fundamenta Math-
ematicae 31, 103–134.
Troelstra, A.S. & van Dalen, D. (1988), Constructivism in Mathematics, II,
number 123 in ‘Studies in logic’, NorthHolland, Amsterdam.
Veldman, W. (1976), ‘An intuitionistic completeness theorem for intuitionis-
tic predicate logic’, The Journal of Symbolic Logic 41, 159–166.
(received May 20, 1974).

10