Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence (IJCAI-16)

The Inconsistency in Gödel’s Ontological Argument:

A Success Story for AI in Metaphysics

Christoph Benzmüller⇤ Bruno Woltzenlogel Paleo

Freie Universität Berlin & Stanford University Australian National University
c.benzmueller@gmail.com bruno.wp@gmail.com

Abstract on the proof [Fuhrmann, 2016].

This paper discusses the discovery of the incon- The in-depth analysis presented here substantially extends
sistency in Gödel’s ontological argument as a suc- previous computer-assisted studies of Gödel’s ontological
cess story for artificial intelligence. Despite the argument. Similarly to the related work [Benzmüller and
popularity of the argument since the appearance Woltzenlogel-Paleo, 2013a; 2014] the analysis has been con-
of Gödel’s manuscript in the early 1970’s, the in- ducted with automated theorem provers for classical higher-
consistency of the axioms used in the argument re- order logic (HOL; cf. [Andrews, 2014] and the references
mained unnoticed until 2013, when it was detected therein), even though Gödel’s proof is actually formulated
automatically by the higher-order theorem prover in higher-order modal logic (HOML; cf. [Muskens, 2006]
Leo-II. Understanding and verifying the refutation and the references therein). To bridge between the two
generated by the prover turned out to be a time- logics we utilise and further improve the logic embedding
consuming task. Its completion, as reported here, approach [Benzmüller and Paulson, 2013; Benzmüller and
required the reconstruction of the refutation in the Woltzenlogel-Paleo, 2014], which has already been employed
Isabelle proof assistant, and it also led to a novel successfully in preceding related work.
and more efficient way of automating higher-order The main novel contribution reported in this paper is a
modal logic S5 with a universal accessibility rela- detailed analysis (in various modal logics) of the inconsis-
tion. Furthermore, the development of an improved tency of Gödel’s original version of the axioms used in his
syntactical hiding for the utilized logic embedding manuscript [1970]. The extraction, reconstruction and veri-
technique allows the refutation to be presented in fication of an informal, human intuitive argument has been
a human-friendly way, suitable for non-experts in an open problem since the first detection of this inconsis-
the technicalities of higher-order theorem proving. tency by Benzmüller and Woltzenlogel-Paleo [2014] with the
This brings us a step closer to wider adoption of Leo-II prover. The verified refutation (discussed in §4) dis-
logic-based artificial intelligence tools by philoso- plays a surprisingly accessible explanation of the inconsis-
phers. tency, which is philosophically profound and never presented
in the literature. The detection of this inconsistency in com-
bination with the work reported here thus demonstrates that
1 Introduction artificial intelligence systems – particularly higher-order au-
Without exaggeration Kurt Gödel’s ontological argument for tomated theorem provers – are capable of assisting in the dis-
the existence of God [Gödel, 1970; Scott, 1972] is amongst covery and elucidation of new and philosophically relevant
the most discussed formal proofs in modern literature. A rich knowledge.
body of publications – including very recent ones – present, On the technical side, the quest for constructing a com-
discuss, assess, criticize, modify and improve Gödel’s orig- pelling refutation, capable of convincing also human non-
inal work (see e.g. Sobel [2004] and Oppy [2015] and the experts, led to an improvement of the syntax of the embed-
references therein). In philosophy lectures at universities ding of modal logics in Isabelle/HOL (as discussed in §3.2).
the argument is regularly presented as a masterpiece argu- With the new syntax, a (nearly) perfect match between the
ment in metaphysics. Since 2013, when Benzmüller and original pen and paper presentations and our encoding in Is-
Woltzenlogel-Paleo [2013a; 2014] first reported their suc- abelle/HOL is feasible. A more user-friendly syntax, as re-
cessful initial computer-assisted analysis of Gödel’s proof ported here, is clearly an important prerequisite for promot-
and Scott’s variant, their work has received a media repercus- ing the theorem proving technology employed here to a wider
sion on a global scale1 , and numerous bloggers commented community of philosophers, who are not necessarily experts
⇤
This work was supported by the German National Research in automated reasoning or HOL.
Foundation (DFG) under grants BE 2501/9-2 and BE 2501/11-1. Another novel contribution reported here (in §3.1) is the
1
A collection of news articles is available at https://github.com/ implementation of an alternative embedding for the more ef-
FormalTheology/GoedelGod/blob/master/Press/LinksToNews.md fective modal logic S5U , which is based on a universal acces-

936
sibility relation. Our experiments have shown that the new
embedding is more efficient, as the following two previously
open problems can now be solved:
• Automatically proving the final theorem T3 (Necessar-
ily, there exists God), directly from Scott’s [1972] (con-
sistent) axioms alone, without relying on the argument’s
intermediate argumentation steps (i.e., lemmata).
• Automatically verifying, in Isabelle/HOL, the proof of
the modal collapse [Sobel, 1987], which is one of the
most strongly criticized logical consequences of the ar-
gument’s axioms.

1.1 Related Work

First successful applications of theorem proving technology
in metaphysics were reported by Fitelson, Oppenheimer and
Zalta [2007; 2011], who coined the term Computational
Metaphysics for this new research area and employed the
first-order Prover9 [McCune, 2010] in their experiments.
Later on, Rushby [2013] used the proof assistant PVS [Owre
et al., 1992]. Common to both works is a significant amount
of proof-hand-coding work as well as their focus on a non-
modal formalization of St. Anselm’s [1078] simpler and older
ontological argument. In contrast, the greater complexity of
Gödel’s argument requires the formalization and automation
of variants of higher-order and modal logics.

2 A Brief History of the Argument

Figure 1: Full Automation of T3 in S5U ; Consistency of
St. Anselm’s ontological argument [Anselm, 1078] can be
Scott’s Axioms; Automatic Verification of Modal Collapse
regarded as the ancestor of modern ontological arguments
such as Gödel’s. In the millenium between Anselm and
Gödel, many philosophers modified and arguably improved several variants have been proposed [Anderson, 1990; An-
Anselm’s argument. Of particular importance to Gödel was derson and Gettings, 1996; Hájek, 1996; 2001; Hájek, 2002;
the work of Leibniz [Adams, 1995]. Although Gödel’s no- Bjørdal, 1999] attempting to avoid the modal collapse. Many
tion of positive property is not exactly the same as Leibniz’s of these variants omit the crucial conjunct in the definition
notion of perfection, Gödel’s manuscript (Fig. 6a) can be con- of essence as well.3 Opponents of the argument (e.g. Oppy
sidered a translation of Leibniz’s presentation of the argument [1996, p.226–227; 2000, p.364; 2008, p.1068]) have also
into modern modal logic. Gödel discussed his manuscript proposed parodies and other criticisms, referring to variants
with Scott, who shared a slightly di↵erent version with a where the conjunct is omitted.
larger public. Scott’s version of the axioms and definitions, However, as explained here, the extra conjunct is in fact
formalized in Isabelle, is shown in Fig. 1. The main di↵er- crucial. Without it, Gödel’s original axioms are inconsis-
ence to Gödel’s version is an extra conjunct in the definition tent. With it, Scott’s axioms are consistent (cf. Fig. 1 where
of essence (ess). Gödel’s di↵erent definition of essence can the model finder Nitpick [Blanchette and Nipkow, 2010] con-
be seen either in his manuscript (Fig. 6a) or, in more mod- firms consistency).4
ern notation, in the Isabelle formalization shown in Fig. 5.
For Scott, an essential property of an individual must be pos-
sessed by him/her. For Gödel, this is not required. 3 Automating HOML in HOL
Gödel’s omission has been considered inessential and Logic textbooks commonly utilize higher-order logic in an
merely an oversight by many. For instance, Hazen [1998, informal/semi-formal way as a meta-language to introduce
p.365] states that “Gödel left this clause out [. . . ] but this the syntax and the semantics of object logics of interest, in
appears to have been an oversight”. For more than four which reasoning problems in concrete application domains
decades, its serious consequences remained unnoticed, de- can be modeled and solved with pen and paper. In fact, this
spite numerous analysis and criticisms of the argument. Es-
pecially since the discovery by Sobel [1987] that modal col- 3
As these variants also change other axioms, on which the in-
lapse (MC)2 is entailed by Gödel’s (or also Scott’s) axioms, consistency of Gödel’s axioms depends, it is not necessarily the case
that these variants are also inconsistent; they must be analyzed sep-
2 arately.
The modal collapse, ! ⇤ , states that contingent truth im-
4
plies necessary truth; it can be interpreted as everything is pre- In personal communication, Dana Scott confirmed that he was
determined or even there is no free will. unaware at the time that Gödel’s axioms were inconsistent.

937
 embed S5 would be to use the standard translation for K de-
scribed above and to state that r is an equivalence relation,
e.g., by postulating the following axioms:
• Reflexivity: 8x.(r x x)
• Symmetry: 8x.8y.(r x y) ! (r y x)
• Transitivity: 8x.8y.8z.(r x y) ^ (r y z) ! (r x z)
Instead, we consider here an alternative description, that we
call S5U , based on the following condition on r:
• Universality: 8x.8y.(r x y)
It is important to note that ✏S5 ' i↵ ✏S5U ' [Blackburn et
al., 2001], and therefore S5 and S5U are traditionally consid-
ered to be two di↵erent descriptions of the same modal logic.
Nevertheless, S5 and S5U di↵er in the shapes of frames they
admit: S5U only admits complete5 frames, whereas S5 admits
non-complete frames as long as all their components are com-
plete. In other words, in S5U we face one single equivalence
class of possible worlds, while in S5 we may face several dis-
connected equivalence classes. In fact, for this reason, S5U
is considered as metaphysically more appropriate by some
philosophers; cf. [Williamson, 2013, p. 127].
Furthermore, for S5U an improved embedding is possible.
Universality implies that the guarding predicates in the defini-
tions of ⇤ and ^ always hold. Therefore, they can be omitted
Figure 2: Improved Embedding of S5U and the accessibility relation can be dispensed altogether. The
modal operators can then be defined merely as:

approach can also be followed on the computer (using HOL ⇤' ⌘ w.8v.'(v) and ^' ⌘ w.9v.'(v)
as a formal meta-language) for even very challenging object
logics (such as HOML) to enable interactive and automated The new embedding of S5U in Isabelle/HOL is shown
theorem proving with existing theorem provers for HOL. in Fig. 2. With this improved embedding, the final theo-
For a computational analysis of Gödel’s ontological argu- rem T3 (Necessarily, there exists God) can be derived from
ment, the embedding of HOMLs such as K, KB and S5 with Scott’s consistent version of the axioms fully automatically.
various domain conditions (possibilist and actualist quantifi- The fully automatic proof has been generated (in about 2.5
cation) is required. This idea has been successfully employed seconds) by the theorem prover Leo-II [Benzmüller et al.,
in related work [Benzmüller and Woltzenlogel-Paleo, 2014]. 2015] and subsequently verified in the proof assistant Is-
The embedding of HOML is in fact straightforward. Formu- abelle/HOL [Nipkow et al., 2002], as shown in Fig. 1. The
las in HOML are lifted, i.e., converted into predicates over collaboration between the two systems has been orchestrated
worlds, which are themselves explicitly represented as terms. by Isabelle’s Sledgehammer tool [Blanchette et al., 2013].
The logical constants of HOML are translated to HOL terms With the embedding used by Benzmüller and
in such a way that, for instance, ⇤' and ^' (relative to a Woltzenlogel-Paleo [2014], the provers still had to be
current world wo ) are mapped, respectively, to the HOL for- given the intermediate theorem T2 and the corollary C in
mulas 8w.(rw0 w)!('w) and 9w.(rw0 w) ^ ('w). This form order to manage to prove T3.
of embedding is precisely the well-known standard transla- Another evidence that the new embedding provides a sig-
tion [Ohlbach, 1991], which is here intra-logically realized nificant performance boost is the successful automatic ver-
— and extended for quantifiers — in HOL by stating a set of ification in Isabelle/HOL (with its automatic tactic Meson)
equations defining the logical constants (Fig. 2). The result- of the modal collapse [Sobel, 1987], which is one of the
ing object logic is the HOML K with rigid terms and constant most strongly criticized ‘side-e↵ects’ of Gödel’s and Scott’s
domains (possibilist quantifiers). Other logics (e.g. KB, S5) variants of the proof. In previous work [2014] the modal
are embedded by adding axioms that restrict the accessibility collapse has been proven by the higher-order provers Satal-
relation r. Varying domains and actualist quantifiers can be lax [Brown, 2012] and Leo-II, but a fully automatic verifi-
simulated by using an existence predicate to guard the quan- cation in the highly trusted Isabelle/HOL system still failed
tifiers. The embedding approach is, therefore, very flexible. [Benzmüller and Woltzenlogel-Paleo, 2013b]. The success
with the new embedding can be seen in Fig. 1.
3.1 Improved Embedding
The modal logic S5 requires that the accessibility relation be 5
A graph is complete i↵ there is a directed edge connecting every
reflexive, symmetric and transitive. The usual approach to ordered pair of vertices.

938
 4 Intuitive Inconsistency Argument
In the typical workflow during an attempt to prove a con-
Figure 3: Definition of Essence using Old Syntax jecture with a theorem prover, it is customary to check the
consistency of the axioms first. For if the axioms are in-
consistent, anything (including the conjecture) would be triv-
3.2 Improved Syntax in Isabelle ially derivable in classical logic (ex falso quodlibet). Sur-
Wider adoption of HOL theorem proving technology for rea- prisingly, when this routine check was performed on Gödel’s
soning about and within embedded object logics, especially axioms [Benzmüller and Woltzenlogel-Paleo, 2014], the Leo-
among non-expert users, is still hindered by the gap between II prover claimed that the axioms were inconsistent. Un-
the syntax used by people, when they write logical formu- fortunately, the refutation generated by Leo-II was barely
las with pen and paper, and the syntax used by HOL theorem human-readable. The text file was 153 lines6 long and used
provers. Even when the syntax of the underlying higher-order machine-oriented calculus (higher-order resolution [Sultana
system is elegant (as is the case in Isabelle/HOL), the em- and Benzmüller, 2013]) and syntax (TPTP THF [Sutcli↵e and
bedding of HOML into HOL may easily expose details of Benzmüller, 2010]). Part of the file is displayed in Fig. 4.
HOL that may be uncommon to the user, disturbing his/her Although Leo-II’s resolution refutation is not easy to read
experience while using the system. To illustrate this point, for humans, it did contain relevant hints to the importance
Fig. 3 shows how the definition of essence looked like in of the empty property x.? (also denoted ;, as in HOL it is
previous work [Benzmüller and Woltzenlogel-Paleo, 2013b], customary to think of unary predicates as sets). Note that
where advanced syntax-sugaring features were not used. It the terms for the empty property7 ( x.?) and for the property
looks notably higher-order, and its style di↵ers significantly of self-di↵erence ( x.x , x) have identical denotations in a
from the common style seen in works on modal logics and logic setting with full functional and Boolean extensionality
the ontological argument. The following specific issues can as given here. Nevertheless, some philosophers8 may actually
be enumerated: prefer the use of self-di↵erence over the empty property in
1. -abstractions, which are typically a HOL feature, ap- the analysis below. However, for the proof to go through it
pear explicitly in places where they did not need to is irrelevant which notion we use and the reader may simply
in a pure HOML formulation (cf. Gödel’s manuscript, replace the empty property by self-di↵erence.
Fig. 6a).
4.1 Informal Argument
2. Quantifiers appear as higher-order defined constants,
and not as binders. This forces the user to read (and Based on the hints found in Leo-II’s refutation, we conceived
write) formulas of the form 8( x.A(x)) instead of the the following informal explanation for the inconsistency of
more common 8x.A(x). Gödel’s axioms:
3. The lifted modal connectives are represented by prefix- 1. From Gödel’s definition of essence
ing the letter “m” (e.g. m^ and m!). The prefix disturbs ( ess x $ 8 ( (x)!⇤8y( (y)! (y)))) it follows
the user, as it constantly reminds him/her that there is that the empty property (or self-di↵erence) is an essence
something unusual about the modal connectives. of every individual (Empty Essence Lemma):
4. Higher-order parenthesis conventions for the application 8x (; ess x)
of a predicate to a term are used. Instead of reading (y),
as he/she would expect, he/she has to read ( y). Outside 2. From theorem T1 (Positive properties are possibly exem-
niche areas in computer science, the former syntax is plified: 8 [P( )!^9x (x)]) and axiom A5 (“necessary
more widely known than the latter. existence” is a positive property: P(NE) ), it follows that
In the embedding presented here, in Fig. 2, advanced NE is possibly exemplified:
syntax-sugaring e↵ects provided by Isabelle were used to ^9x[NE(x)]
prevent issues as those enumerated above. The possibility
to define boldface connectives allows us to drop the prefix; 3. Expanding the definition of “necessary existence”
“binder” annotations enable modal quantifiers to be used in (NE(x) ⌘ 8 [ ess x!⇤9y (y)]), the following is ob-
the standard binding way and reduce the need for explicit tained:
lambda abstractions; and a careful choice of priorities for ^9x[8'[' ess x!⇤9y['(y)]]]
infix connectives gives the parenthesis conventions that are
more familiar to the user. As desired, the definition of essence 4. The sentence above holds for all ' and thus, in particular,
in Fig. 1 is undeniably more immediately recognizable and for the empty property (or self-di↵erence):
comprehensible than the definition in Fig. 3. The embedding
technique is now completely transparent to the user. ^9x[; ess x!⇤9y[;(y)]]
The syntax improvements described here render the 6
Long lines with an average of 184 characters per line.
computer-assisted analysis of ontological arguments accessi- 7
An additional lambda abstraction occurs in the empty property
ble to a wider audience and ease the adoption of logic-based in Leo-II’s proof (and also in the reconstruction in Isabelle) because
artificial intelligence tools by philosophers interested in top- the embedding approach lifts the boolean type o to ◆!o.
ics where modal logic reasoning is required. 8
Private communication with André Fuhrmann.

939
Figure 4: Lines 115–120 of Leo-II’s refutation. Primitive substitutions (e.g. with the empty property) are highlighted. In the
red part (see ), property variable SV8 has been instantiated with the SV16µ . SV17◆ .?, i.e., the (lifted) empty property.

5. By the Empty Essence Lemma, the antecedent of the im-

plication above is valid. Therefore, the sentence above
entails:
^9x[⇤9y[;(y)]]
6. By definition of ;:
^9x[⇤?]
7. As the existential quantifier is binding no variable within
its scope, the sentence is equi-valid with:
^⇤?
8. To see that the sentence above is contradictory, we may
reason semantically, thinking of possible worlds. If w0
is the arbitrary current world, the ^ operator forces the
existence of a world w accessible from w0 such that ⇤?
is true in w. But ⇤? can only be true in w, if there is no
world w0 accessible from w. In logics with a reflexive or
symmetric accessibility relation (e.g. KB), it is easy to
see that there must be a world w0 accessible from w: ei-
ther w0 itself, in case of a reflexive relation, or w0 , in case
of a symmetric relation. In fact, even in K, with no ac-
cessibility condition, there must be a world w0 accessible
from w. The reason is that ^⇤? should be valid (true in
all worlds). Therefore, it is true in w as well, where the Figure 5: Inconsistency of Gödel’s Axioms in HOML K ver-
existence of an accessible world w0 is forced by the ^ ified in Isabelle/HOL
operator. As a model for ^⇤? (which is a consequence
of Gödel’s axioms) cannot be built, Gödel’s axioms are
inconsistent. automate the inessential parts (cf. Fig. 5). The essential use
Interestingly, the refutation automatically generated by of the Empty Essence Lemma, on the other hand, is explicitly
Leo-II uses a symmetric accessibility relation, and thus re- stated, to ensure that Isabelle is reconstructing the same argu-
quires the modal logic KB. The informal, human-constructed ment. In fact, without the help of this lemma, Metis is still
refutation described above, on the other hand, requires only not strong enough to refute Gödel’s axioms.
the weaker modal logic K. In our experiments Leo-II (like all
other HOL provers) was still too weak to automatically prove 4.3 Mapping the Inconsistency to Gödel
the inconsistency already in logic K. Hence, this remains an The inconsistency verified in Fig. 5 follows from the defi-
open problem for automated theorem provers. nition of essence (ess), the definition of necessary existence
4.2 Argument Reconstruction in Isabelle (NE), the axioms A1a and A2 (which entail theorem T1), and
axiom A5. It remains to show that these ingredients are actu-
To verify the correctness of the informal argument explained ally present in Gödel’s manuscript in Fig. 6a.
above, it was reconstructed in Isabelle/HOL, using Metis9 to This can be easily seen: Axiom A1a in Fig. 5 is implied by
9
Metis, unlike external provers such as Leo-II or Satallax, con- Axiom Ax2 and the highlighted footnote remark in Fig. 6a.
structs proofs in Isabelle’s highly trusted kernel calculus. Axioms A2 and A5 in Fig. 5 correspond to Ax4 and Ax3 in

940
(a) Gödel’s manuscript, with mutually inconsistent axioms and definitions highlighted (with permis- (b) Inconsistency in HOML S5U
sion from the Kurt Gödel Papers, Shelby White and Leon Levy Archives Center, Princeton, NJ, USA,
on deposit at Princeton University)

Figure 6: The inconsistency in Gödel’s manuscript has been detected and verified by HOL ATPs

Fig. 6a. The definitions of essence and necessary existence On the other hand, there is still room for many pragmat-
are easy to identify. Therefore, the verified inconsistency ical improvements in Isabelle; just one example: in default
from Fig. 5 does apply to Gödel’s original manuscript. setting, Sledgehammer does not immediately inform the user
when a proof has been found and instead silently first ex-
4.4 Inconsistency of Gödel’s Axioms in S5U ecutes a series of time-consuming proof analysis processes
Isabelle/HOL’s Sledgehammer tool, which orchestrates calls (e.g. its dependency minimization), before it eventually re-
to external provers such as Leo-II, still fails to detect the in- ports success. For Gödel’s theorem T3 (Necessarily, there
consistency of Gödel’s axioms in the standard embedding of exists God), for example, this phase of silence takes several
S5, while a direct modeling of the problem in TPTP THF syn- minutes — during which the user might actually give up on
tax in combination with a direct call of Leo-II succeeded. In the proof attempt — even though Leo-II already reported suc-
other words, without independent experiments with no me- cess to Sledgehammer after 2.5 seconds.
diation through Sledgehammer, the inconsistency would not More importantly, our work reveals a challenge for auto-
have been detected. mated reasoning: the (so far partially manual) extraction of
On the other hand (and further confirming the claims from an informal argument from a formal proof. Without accom-
§3.1), the reconstruction in Isabelle/HOL with the improved panying human-understandable explanations, the proofs gen-
embedding for S5U was more efficient: the inconsistency erated by provers such as Leo-II or Metis, will presumably
could be detected by Leo-II also when called via Sledgeham- be only of limited value for philosophers, for whom intuitive
mer. Moreover, the result could subsequently be verified with arguments remain crucial for the acceptance of novel results.
Metis even without the Empty Essence Lemma (cf. Fig. 6b). Another open problem that we solved in this paper is a fully
automatic proof of T3 directly from Scott’s axioms. Again,
5 Conclusion this proof was contributed by Leo-II. This has become pos-
sible only after we provided a more efficient embedding for
The axioms and definitions in Gödel’s manuscript are incon- HOML S5U (instead of S5) in HOL.
sistent; this was detected automatically by the prover Leo-II. Both the automated detection of the inconsistency in
Here we presented a rational reconstruction and verification Gödel’s axioms and the fully automatic proof of T3 from
of the inconsistency argument in Isabelle/HOL. This argu- Scott’s axioms demonstrate the potential of our AI technol-
ment is valid in all normal HOMLs including base logic K. ogy for philosophy: this technology is, in its current state of
We have also presented several technical improvements re- development, already capable of contributing novel results to
garding the semantic embedding approach. In particular, we metaphysics and to conduct reasoning steps at granularity-
have achieved a nearly perfect match between pen and paper levels beyond common human capabilities.
presentations in HOML and the syntax in Isabelle/HOL. As a
result, the embedding of HOML in HOL is now fully trans- Acknowledgments: We thank Chad Brown, who contributed
parent, more user-friendly and ready for wider adoption. to the rational reconstruction of the inconsistency argument.

941
References [Hájek, 1996] P. Hájek. Magari and others on Gödels ontological
[Adams, 1995] R.M. Adams. Introductory note to *1970. In Kurt proof. In A. Ursini and P. Agliano, editors, Logic and algebra,
Gödel: Collected Works Vol. 3: Unpubl. Essays and Letters. Ox- page 125135. Dekker, New York etc., 1996.
ford Univ. Press, 1995. [Hájek, 2001] P. Hájek. Der Mathematiker und die Frage der Ex-
[Anderson and Gettings, 1996] A.C. Anderson and M. Gettings. istenz Gottes. In B. Buldt et al., editor, Kurt Gödel. Wahrheit
Gödel ontological proof revisited. In Gödel’96: Logical Founda- und Beweisbarkeit, pages 325–336. bv & hpt, Wien, 2001. ISBN
tions of Mathematics, Computer Science, and Physics: Lecture 3-209-03835-X.
Notes in Logic 6, pages 167–172. Springer, 1996. [Hájek, 2002] P. Hájek. A new small emendation of Gödel’s onto-
[Anderson, 1990] C.A. Anderson. Some emendations of Gödel’s logical proof. Studia Logica, 71(2):149–164, 2002.
ontological proof. Faith and Philosophy, 7(3), 1990. [Hazen, 1998] A.P. Hazen. On Gödel’s ontological proof. Aus-
tralasian Journal of Philosophy, 76:361–377, 1998.
[Andrews, 2014] P.B. Andrews. Church’s type theory. In E.N.
Zalta, editor, The Stanford Encyclopedia of Philosophy. Spring [McCune, 2010] W. McCune. Prover9 and Mace4 (2005–2010).
2014 edition, 2014. http://www.cs.unm.edu/˜mccune/prover9/, 2010.
[Anselm, 1078] St. Anselm. Proslogion. In M. Charlesworth, edi- [Muskens, 2006] R. Muskens. Higher Order Modal Logic. In
tor, St. Anselm’s Proslogion. Oxford:OUP, 1078. Republished in P. Blackburn et al., editor, Handbook of Modal Logic, Studies
1965. in Logic and Practical Reasoning, pages 621–653. Elsevier, Dor-
drecht, 2006.
[Benzmüller and Paulson, 2013] C. Benzmüller and L.C. Paulson.
Quantified multimodal logics in simple type theory. Logica Uni- [Nipkow et al., 2002] T. Nipkow, L. Paulson, and M. Wenzel. Is-
versalis, 7(1):7–20, 2013. abelle/HOL: A Proof Assistant for Higher-Order Logic. Number
2283 in LNCS. Springer, 2002.
[Benzmüller and Woltzenlogel-Paleo, 2013a] C. Benzmüller and
[Ohlbach, 1991] H.J. Ohlbach. Semantics-based translation meth-
B. Woltzenlogel-Paleo. Formalization, mechanization and au-
tomation of Gödel’s proof of God’s existence. arXiv:1308.4526, ods for modal logics. J. Log. Comput., 1(5):691–746, 1991.
2013. Preprint available as arXiv:1308.4526. [Oppenheimer and Zalta, 2011] P.E. Oppenheimer and E.N. Zalta.
[Benzmüller and Woltzenlogel-Paleo, 2013b] C. Benzmüller and A computationally-discovered simplification of the ontological
B. Woltzenlogel-Paleo. Gödel’s God in Isabelle/HOL. Archive argument. Australasian J. of Philosophy, 89(2):333–349, 2011.
of Formal Proofs, 2013, 2013. [Oppy, 1996] G. Oppy. Gödelian ontological arguments. Analysis,
[Benzmüller and Woltzenlogel-Paleo, 2014] C. Benzmüller and 56(4):226–230, 1996.
B. Woltzenlogel-Paleo. Automating Gödel’s ontological proof of [Oppy, 2000] G. Oppy. Response to Gettings. Analysis, 60(4):363–
God’s existence with higher-order automated theorem provers. 367, 2000.
In ECAI 2014, volume 263 of Frontiers in Artificial Intelligence [Oppy, 2008] G. Oppy. Higher-order ontological arguments. Phi-
and Applications, pages 93 – 98. IOS Press, 2014. losophy Compass, 3(5):1066–1078, 2008.
[Benzmüller et al., 2015] Christoph Benzmüller, Lawrence C. [Oppy, 2015] G. Oppy. Ontological arguments. In Edward N. Zalta,
Paulson, Nik Sultana, and Frank Theiß. The higher-order prover editor, The Stanford Encyclopedia of Philosophy. Spring 2015
LEO-II. J. of Automated Reasoning, 55(4):389–404, 2015. edition, 2015.
[Bjørdal, 1999] F. Bjørdal. Understanding Gödels ontological argu- [Owre et al., 1992] S. Owre, J. M. Rushby, and N. Shankar. PVS: A
ment. In T. Childers, editor, The Logica Yearbook 1998. Filosofia, prototype verification system. In Deepak Kapur, editor, CADE,
1999. volume 607 of LNAI, pages 748–752, Saratoga, NY, jun 1992.
[Blackburn et al., 2001] P. Blackburn, M. de Rijke, and Y. Venema. Springer.
Modal Logic. Cambridge University Press, 2001. [Rushby, 2013] J. Rushby. The ontological argument in PVS. In
[Blanchette and Nipkow, 2010] J.C. Blanchette and T. Nipkow. Proc. of CAV Workshop “Fun With Formal Methods”, St. Peters-
Nitpick: A counterexample generator for higher-order logic burg, Russia, 2013.
based on a relational model finder. In ITP 2010, number 6172 [Scott, 1972] D. Scott. Appx. B: Notes in Dana Scott’s Hand, pages
in LNCS, pages 131–146. Springer, 2010. 145–146. In Sobel [2004], 1972.
[Blanchette et al., 2013] J.C. Blanchette, S. Böhme, and L.C. Paul- [Sobel, 1987] J.H. Sobel. Gödel’s ontological proof. In On Being
son. Extending Sledgehammer with SMT solvers. J. of Auto- and Saying. Essays for Richard Cartwright, pages 241–261. MIT
mated Reasoning, 51(1):109–128, 2013. Press, 1987.
[Brown, 2012] C.E. Brown. Satallax: An automated higher-order [Sobel, 2004] J.H. Sobel. Logic and Theism: Arguments for and
prover. In IJCAR 2012, number 7364 in LNAI, pages 111 – 117. Against Beliefs in God. Cambridge U. Press, 2004.
Springer, 2012. [Sultana and Benzmüller, 2013] N. Sultana and C. Benzmüller. Un-
[Fitelson and Zalta, 2007] Branden Fitelson and Edward N. Zalta. derstanding LEO-II’s proofs. In Konstantin Korovin, Stephan
Steps toward a computational metaphysics. J. Philosophical Schulz, and Eugenia Ternovska, editors, IWIL 2012, volume 22
Logic, 36(2):227–247, 2007. of EPiC Series, pages 33–52, Merida, Venezuela, 2013. Easy-
Chair.
[Fuhrmann, 2016] A. Fuhrmann. Blogging Gödel: His ontological
[Sutcli↵e and Benzmüller, 2010] G. Sutcli↵e and C. Benzmüller.
argument in the public eye. In K. Świetorzecka, editor, Forth-
coming in Gödels Ontological Argument - History, Modifications, Automated reasoning in higher-order logic using the TPTP THF
and Controversies. 2016. infrastructure. J. of Formalized Reasoning, 3(1):1–27, 2010.
[Williamson, 2013] T. Williamson. Modal Logic as Metaphysics.
[Gödel, 1970] K. Gödel. Appx. A: Notes in Kurt Gödel’s Hand,
Oxford:OUP, 2013.
pages 144–145. In Sobel [2004], 1970.

942