Falsification-aware Semantics for CTL and Its Inconsistency-tolerant
Subsystem: Towards Falsification-aware Model Checking
Norihiro Kamide and Seidai Kanbe
Teikyo University, Faculty of Science and Engineering, Department of Information and Electronic Engineering,
Toyosatodai 1-1, Utsunomiya-shi, Tochigi 320-8551, Japan
Keywords:
Computation Tree Logic, Inconsistency-tolerant Computation Tree Logic, Falsification-aware Kripke-style
Semantics, Falsification-aware Model Checking.
Abstract:
This study introduces two falsification-aware Kripke-style semantics for computation tree logic (CTL). The
equivalences among the proposed falsification-aware Kripke-style semantics and the standard Kripke-style
semantics for CTL are proven. Furthermore, a new logic, inconsistency-tolerant CTL (ICTL) is semantically
defined and obtained from the proposed falsification-aware Kripke-style semantics for CTL by deleting a char-
acteristic condition on the labeling function of the semantics. Because ICTL is regarded as an inconsistency-
tolerant and many-valued logic, the proposed semantic framework for CTL and ICTL is regarded as a unified
framework for combining and generalizing the standard, inconsistency-tolerant, and many-valued semantic
frameworks. This unified semantic framework is useful for generalized model checking, referred to here as
falsification-aware model checking.
1 INTRODUCTION
Computation tree logic (CTL) (Clarke and Emer-
son, 1981) is known to be one of the most use-
ful temporal logics for model checking (Clarke and
Emerson, 1981; Clarke et al., 2018), which is a
computer-assisted method used to verify concurrent
systems that can be modeled by state-transition sys-
tems. Some inconsistency-tolerant variants of CTL,
referred to as inconsistency-tolerant CTLs, paracon-
sistent CTLs, and many-valued CTLs, have been
developed as a theoretical basis for inconsistency-
tolerant (paraconsistent and many-valued) model
checking (Easterbrook and Chechik, 2001; Chen and
Wu, 2006). Some inconsistency-tolerant CTLs have
falsification-aware Kripke-style semantics, which are
capable of representing the explicit falsification of a
given negated formula and are appropriate for speci-
fying and verifying inconsistency-tolerant reasoning.
For more information on inconsistency-tolerant tem-
poral logics and their model checking applications,
see (Easterbrook and Chechik, 2001; Chen and Wu,
2006; Kamide, 2006; Kamide and Wansing, 2011;
Kamide and Kaneiwa, 2010; Kaneiwa and Kamide,
2011; Kamide, 2015; Kamide and Koizumi, 2016;
Kamide and Endo, 2018).
In this study, we first introduce two new
falsification-aware Kripke-style semantics for CTL:
falsification-aware normal Kripke-style semantics
and falsification-aware dual Kripke-style semantics.
We then prove the equivalences among the proposed
falsification-aware Kripke-style semantics and the
standard Kripke-style semantics for CTL. Second, we
also semantically define a new logic, inconsistency-
tolerant CTL (ICTL), which is obtained from the
proposed falsification-aware normal and dual Kripke-
style semantics for CTL by deleting a characteris-
tic condition on the labeling function of the seman-
tics. Because ICTL is regarded as an inconsistency-
tolerant and many-valued logic, the proposed seman-
tic framework for CTL and ICTL is regarded as a uni-
fied framework for combining and generalizing the
standard, inconsistency-tolerant, and many-valued se-
mantic frameworks. This unified semantic framework
is useful for combined and generalized model check-
ing, referred to here as falsification-aware model
checking.
Next, we explain the motivation for develop-
ing falsification-aware Kripke-style semantics for
CTL and ICTL. An adequate representation of
falsification-aware reasoning is considered a major
concern in philosophical logic (Horn and Wansing,
2017). It was suggested in (Kamide, 2021) that se-
242
Kamide, N. and Kanbe, S.
Falsification-aware Semantics for CTL and Its Inconsistency-tolerant Subsystem: Towards Falsification-aware Model Checking.
DOI: 10.5220/0010803700003116
In Proceedings of the 14th International Conference on Agents and Artificial Intelligence (ICAART 2022) - Volume 3, pages 242-252
ISBN: 978-989-758-547-0; ISSN: 2184-433X
Copyright
c
2022 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
mantics and/or proof systems for a logic can be con-
sidered falsification-aware if they are capable of pro-
viding (or representing) the direct (or explicit) fal-
sifications or refutations of given negated formulas
(except for negated atomic formulas). Furthermore,
falsification-aware Kripke-style semantics are suit-
able for the theoretical basis of model checking be-
cause falsification plays a critical role in obtaining
counterexample traces for the underlying object spec-
ifications in model checking. In fact, a falsification-
aware technique for explicitly dividing falsification
(or refutation) and verification (or justification) is of-
ten used for model checking (Gurfinkel et al., 2006).
A counterexample-guided abstraction and refinement
technique (Clarke et al., 2003) for model check-
ing, which is based on falsification-aware Kripke-
style semantics in temporal logics, is also an exam-
ple of a useful technique in model checking. How-
ever, the standard Kripke-style semantics for CTL is
not falsification-aware. Thus, we intend to develop
falsification-aware Kripke-style semantics for CTL to
obtain a concrete logical foundation of falsification-
aware model checking, which is defined as model
checking based on falsification-aware semantics.
The proposed falsification-aware normal and dual
Kripke-style semantics for CTL are constructed on
the basis of the idea of falsification-aware settings for
Nelson’s inconsistency-tolerant (or paraconsistent)
four-valued logic N4 (Almukdad and Nelson, 1984;
Nelson, 1949). Moreover, they are regarded as gen-
eralizations of the previously proposed falsification-
aware Kripke-style semantics for some inconsistency-
tolerant CTLs. The proposed falsification-aware
Kripke-style semantics for CTL have no standard con-
dition “|= ¬α iff 6|= α” with respect to the satisfaction
relation |= and the classical negation connective ¬.
This standard condition represents the transformation
of a negated formula to a non-negated formula (i.e.,
it is not falsification-aware). Instead of this condi-
tion, the falsification-aware dual Kripke-style seman-
tics for CTL has two clearly divided satisfaction re-
lations |=
+
and |=
, which explicitly represent ver-
ification (or justification) and falsification (or refuta-
tion), respectively. Thus, we can obtain the appropri-
ate falsification-aware reasoning with |=
. The pro-
posed falsification-aware Kripke-style semantics have
the merit of simply defining a natural inconsistency-
tolerant CTL as a subsystem of CTL. In fact, the
inconsistency-tolerant temporal logic ICTL proposed
here can be defined in a modular way from the
falsification-aware Kripke-style semantics for CTL
by deleting only one mapping condition for the label-
ing function of the semantics. This type of simple def-
inition cannot be obtained using the standard Kripke-
style semantics for CTL. Thus, we can generalize
the framework and concept of previously proposed
inconsistency-tolerant (or paraconsistent) temporal
logics and inconsistency-tolerant (paraconsistent or
many-valued) model checking by using the proposed
falsification-aware Kripke-style semantics and the
new concept of falsification-aware model checking.
In what follows, we discuss existing inconsistency-
tolerant temporal logics and inconsistency-tolerant
model checking.
Compared to other non-classical logics,
inconsistency-tolerant (or paraconsistent) logics
including inconsistency-tolerant temporal logics have
no axiom of explosion, (α∧¬α)β. Hence, they
can be appropriately used in inconsistency-tolerant
reasoning (Priest and Tanaka, 2009; da Costa et al.,
1995; Wansing, 1993). For example, the following
undesirable scenario is considered. The formula
(s(x)∧¬s(x))d(x), which is an instance of the
axiom of explosion, is valid for any symptom s and
disease d, where ¬s(x) represents “a person x does
not have a symptom s and d(x) represents “a person
x suffers from a disease d. The inconsistent scenario
written as clinical-depression(Maria)∧¬clinical-
depression(Maria) will inevitably arise from
the uncertain definition of clinical depression (or
melancholia). The statement “Maria has clinical de-
pression” can be judged as true or false on the basis of
the perception of different pathologists. In this case,
the formula (clinical-depression(Maria)∧¬clinical-
depression(Maria))cancer(Maria) is valid in
classical logic and standard temporal logics (as an
inconsistency that has an undesirable consequence)
but invalid in inconsistency-tolerant logics (as these
logics are inconsistency-tolerant). ICTL is also
regarded as an inconsistency-tolerant logic, although
CTL is not.
Several inconsistency-tolerant temporal logics
have been developed. For example, multi-valued
computation tree logic, referred to as χCTL, was de-
veloped by Easterbrook and Chechik (Easterbrook
and Chechik, 2001) to realize multi-valued model
checking, which is regarded as the first considera-
tion for inconsistency-tolerant model checking. Qua-
siclassical temporal logic, referred to as QCTL,
was developed by Chen and Wu (Chen and Wu,
2006) to handle inconsistent concurrent systems us-
ing inconsistency-tolerant model checking. Para-
consistent full computation tree logic, referred to as
4CTL
, was proposed by Kamide (Kamide, 2006)
for an expressive base logic for inconsistency-tolerant
model checking. A paraconsistent linear-time tempo-
ral logic, referred to as PLTL, based on a Gentzen-
style sequent calculus was developed by Kamide
Falsification-aware Semantics for CTL and Its Inconsistency-tolerant Subsystem: Towards Falsification-aware Model Checking
243
and Wansing (Kamide and Wansing, 2011) to ana-
lyze philosophical issues. A paraconsistent compu-
tation tree logic, referred to as PCTL, was devel-
oped by Kamide and Kaneiwa (Kamide and Kaneiwa,
2010; Kaneiwa and Kamide, 2011) to realize ef-
ficient inconsistency-tolerant CTL-model checking.
Another paraconsistent CTL, referred to as pCTL, and
paraconsistent linear-time temporal logic, referred
to as pLTL, were developed by Kamide and Endo
(Kamide and Endo, 2018; Kamide and Endo, 2019)
to realize simple and efficient inconsistency-tolerant
model checking. We remark that the name pCTL
for paraconsistent CTL is rather confusing because
the same name has also been used for probabilistic
CTL (Aziz et al., 1995; Bianco and de Alfaro, 1995),
and this probabilistic temporal logic pCTL has been
used for probabilistic model checking. Probabilistic
temporal logic was developed by Aziz et al. (Aziz
et al., 1995) and Bianco and de Alfaro (Bianco and
de Alfaro, 1995) to verify probabilistic concurrent
systems. In the present study, ICTL is regarded as or
closely related to the classical negation-free subsys-
tems of PCTL and the inconsistency-tolerant tempo-
ral logic pCTL. Namely, ICTL is regarded as a purely
inconsistency-tolerant subsystem of PCTL and pCTL.
Several extensions of inconsistency-tolerant tem-
poral logics have been developed. For example,
a sequence-indexed paraconsistent computation tree
logic, referred to as SPCTL, was developed by
Kamide (Kamide, 2015) by extending CTL with the
addition of both a paraconsistent negation connective
and a sequence modal operator. SPCTL was used to
verify clinical reasoning with inconsistent and hier-
archical information. An inconsistency-tolerant (or
paraconsistent) probabilistic computation tree logic,
referred to as PpCTL, was developed by Kamide
and Koizumi (Kamide and Koizumi, 2015; Kamide
and Koizumi, 2016) to verify stochastic or ran-
domized inconsistent concurrent systems. A loca-
tive inconsistency-tolerant hierarchical probabilistic
computation tree logic, referred to as LIHpCTL, was
developed by Kamide and Bernal (Kamide and Bernal
J.P.A., 2019) as an extension of PpCTL and the hi-
erarchical (or sequential) computation tree logic, re-
ferred to as sCTL, developed in (Kamide and Yano,
2017; Kamide, 2018). An inconsistency-tolerant (or
paraconsistent) hierarchical probabilistic computa-
tion tree logic, referred to as IHpCTL, was devel-
oped by Kamide and Yamamoto (Kamide and Ya-
mamoto, 2021) as a modified version of the location-
operator-free subsystem of LIHpCTL. For more in-
formation on (extended) inconsistency-tolerant tem-
poral logics and their model checking applications,
see (Easterbrook and Chechik, 2001; Chen and Wu,
2006; Kamide, 2006; Kamide and Wansing, 2011;
Kamide and Kaneiwa, 2010; Kaneiwa and Kamide,
2011; Kamide, 2015; Kamide and Koizumi, 2016;
Kamide and Endo, 2018).
The remainder of this paper is organized as fol-
lows. In Section 2, as the preliminaries of this study,
we define the standard syntax and Kripke-style se-
mantics for CTL. In Section 3, we introduce the
falsification-aware normal Kripke-style semantics for
CTL and prove the equivalence between the standard
Kripke-style semantics and the proposed falsification-
aware normal Kripke-style semantics. In Section 4,
we introduce the falsification-aware dual Kripke-style
semantics for CTL and prove the equivalence between
the standard Kripke-style semantics and the proposed
falsification-aware dual Kripke-style semantics. Fur-
ther, we obtain the equivalences among the standard
and two proposed falsification-aware Kripke-style se-
mantics for CTL. In Section 5, we first semantically
define ICTL by deleting a characteristic mapping con-
dition from the falsification-aware normal and dual
Kripke-style semantics for CTL. We then prove the
equivalence between the falsification-aware normal
and dual Kripke-style semantics for ICTL. In Section
6, we conclude this paper and address some remarks
on falsification-aware model checking based on the
proposed and extended falsification-aware Kripke-
style semantics.
2 NORMAL KRIPKE-STYLE
SEMANTICS FOR CTL
Formulas of CTL are constructed from countably
many propositional variables by the logical connec-
tives: (implication), (conjunction), (disjunc-
tion), and ¬ (negation); the temporal operators: X
(next-time), G (globally or any-time in the future), F
(eventually or some-time in the future), U (until), and
R (release); and the path quantifiers: A (all computa-
tion paths) and E (some computation path). We use
an expression α β to denote (αβ)(βα). We
use lower-case letters p,q,... to denote propositional
variables, Greek lower-case letters α, β, ... to denote
formulas, and lower-case letters i, j and k to denote
any natural numbers. We use the symbol ω to repre-
sent the set of natural numbers, the symbol or
to denote the linear order on ω, and the symbol to
denote the equality of symbols.
Definition 2.1. Formulas of CTL are defined by the
following Backus-Naur form, assuming p represents
propositional variables:
α ::= p | α α | α α | αα | ¬α
ICAART 2022 - 14th International Conference on Agents and Artificial Intelligence
244
| AXα | EXα | AGα | EGα | AFα | EFα
| A(αUα) | E(αUα) | A(αRα) | E(αRα).
Definition 2.2 (Normal Kripke-style semantics for
CTL). A structure (S,S
0
,R,L) is called a CTL-model if
1. S is the set of states,
2. S
0
is a set of initial states and S
0
S,
3. R is a binary relation on S such that s S s
0
S [(s,s
0
) R],
4. L is a mapping from S to the power set of a
nonempty set Φ of propositional variables.
A path in a CTL-model is an infinite sequence π =
s
0
,s
1
,s
2
,... of states such that i 0 [(s
i
,s
i+1
) R].
A CTL-satisfaction relation (M,s) |= α for any
formula α, where M is a CTL-model (S,S
0
,R,L) and
s is a state in S, is defined inductively by the following
clauses:
1. for any p Φ, (M, s) |= p iff p L(s),
2. (M,s) |= αβ iff (M,s) |= α and (M, s) |= β,
3. (M,s) |= αβ iff (M,s) |= α or (M, s) |= β,
4. (M,s) |= αβ iff (M,s) 6|= α or (M, s) |= β,
5. (M,s) |= ¬α iff (M,s) 6|= α,
6. (M,s) |= AXα iff s
1
S [(s,s
1
) 6∈ R or (M, s
1
) |= α],
7. (M,s) |= EXα iff s
1
S [(s,s
1
) R and (M,s
1
) |= α],
8. (M,s) |= AGα iff for any paths π s
0
,s
1
,s
2
,... with
s s
0
, and any states s
i
along π, we have (M,s
i
) |= α,
9. (M,s) |= EGα iff there exists a path π s
0
,s
1
,s
2
,...
with s s
0
, and for any states s
i
along π, we have
(M,s
i
) |= α,
10. (M,s) |= AFα iff for any paths π s
0
,s
1
,s
2
,... with s
s
0
, there exists a state s
i
along π such that (M,s
i
) |= α,
11. (M,s) |= EFα iff there exists a path π s
0
,s
1
,s
2
,...
with s s
0
, and for some state s
i
along π, we have
(M,s
i
) |= α,
12. (M,s) |= A(αUβ) iff for any paths π s
0
,s
1
,s
2
,... with
s s
0
, there exists a state s
j
along π such that (M,s
j
) |=
β and 0 k < j (M,s
k
) |= α,
13. (M,s) |= E(αUβ) iff there exists a path π s
0
,s
1
,s
2
,...
with s s
0
, and for some state s
j
along π, we have
(M,s
j
) |= β and 0 k < j (M, s
k
) |= α,
14. (M,s) |= A(αRβ) iff for any paths π s
0
,s
1
,s
2
,... with
s s
0
, and any states s
j
along π, we have (M,s
j
) |= β
or 0 k < j (M, s
k
) |= α,
15. (M,s) |= E(αRβ) iff there exists a path π s
0
,s
1
,s
2
,...
with s s
0
, and for any states s
j
along π, we have
(M,s
j
) |= β or 0 k < j (M, s
k
) |= α.
A formula α is called valid in CTL if (M,s) |= α
holds for any CTL-model M := (S,S
0
,R,L), any s S,
and any CTL-satisfaction relation |= on M.
3 FALSIFICATION-AWARE
NORMAL KRIPKE-STYLE
SEMANTICS FOR CTL
A falsification-aware normal Kripke-style semantics
for CTL is defined as follows.
Definition 3.1 (Falsification-aware normal Krip-
ke-style semantics for CTL). Let Φ be a non-empty
set of propositional variables and Φ
¬
be the set
p | p Φ} of negated propositional variables.
A structure (S,S
0
,R,L
?
) is called a falsification-
aware normal CTL-model if
1. S is the set of states,
2. S
0
is a set of initial states and S
0
S,
3. R is a binary relation on S such that s S s
0
S [(s,s
0
) R],
4. L
?
is a mapping from S to the power set of Φ Φ
¬
such that for any p Φ and any s S, p L
?
(s)
iff ¬p 6∈ L
?
(s).
A path in a falsification-aware normal CTL-model
is defined in the same way as that for CTL-model.
A falsification-aware normal CTL-satisfaction re-
lation (M,s) |=
?
α for any formula α, where M is
a falsification-aware normal CTL-model (S,S
0
,R,L
?
)
and s is a state in S, is defined inductively by the fol-
lowing clauses:
1. for any p Φ, (M, s) |=
?
p iff p L
?
(s),
2. (M,s) |=
?
αβ iff (M,s) |=
?
α and (M,s) |=
?
β,
3. (M,s) |=
?
αβ iff (M,s) |=
?
α or (M,s) |=
?
β,
4. (M,s) |=
?
αβ iff (M,s) 6|=
?
α or (M,s) |=
?
β,
5. (M,s) |=
?
AXα iff s
1
S [(s,s
1
) 6∈ R or (M,s
1
) |=
?
α],
6. (M,s) |=
?
EXα iff s
1
S [(s, s
1
) R and (M,s
1
) |=
?
α],
7. (M,s) |=
?
AGα iff for any paths π s
0
,s
1
,s
2
,... with
s s
0
, and any states s
i
along π, we have (M,s
i
) |=
?
α,
8. (M,s) |=
?
EGα iff there exists a path π s
0
,s
1
,s
2
,...
with s s
0
, and for any states s
i
along π, we have
(M,s
i
) |=
?
α,
9. (M,s) |=
?
AFα iff for any paths π s
0
,s
1
,s
2
,...
with s s
0
, there exists a state s
i
along π such that
(M,s
i
) |=
?
α,
10. (M,s) |=
?
EFα iff there exists a path π s
0
,s
1
,s
2
,...
with s s
0
, and for some state s
i
along π, we have
(M,s
i
) |=
?
α,
11. (M,s) |=
?
A(αUβ) iff for any paths π s
0
,s
1
,s
2
,...
with s s
0
, there exists a state s
j
along π such that
(M,s
j
) |=
?
β and 0 k < j (M,s
k
) |=
?
α,
12. (M,s) |=
?
E(αUβ) iff there exists a path π
s
0
,s
1
,s
2
,... with s s
0
, and for some state s
j
along π,
we have (M,s
j
) |=
?
β and 0 k < j (M,s
k
) |=
?
α,
13. (M,s) |=
?
A(αRβ) iff for any paths π s
0
,s
1
,s
2
,...
with s s
0
, and any states s
j
along π, we have
(M,s
j
) |=
?
β or 0 k < j (M,s
k
) |=
?
α,
Falsification-aware Semantics for CTL and Its Inconsistency-tolerant Subsystem: Towards Falsification-aware Model Checking
245
14. (M,s) |=
?
E(αRβ) iff there exists a path π
s
0
,s
1
,s
2
,... with s s
0
, and for any states s
j
along π,
we have (M,s
j
) |=
?
β or 0 k < j (M,s
k
) |=
?
α,
15. for any ¬p Φ
¬
, (M,s) |=
?
¬p iff ¬p L
?
(s),
16. (M,s) |=
?
¬¬α iff (M,s) |=
?
α,
17. (M,s) |=
?
¬(αβ) iff (M,s) |=
?
¬α or (M,s) |=
?
¬β,
18. (M,s) |=
?
¬(αβ) iff (M,s) |=
?
¬α and (M, s) |=
?
¬β,
19. (M,s) |=
?
¬(αβ) iff (M,s) |=
?
α and (M,s) |=
?
¬β,
20. (M,s) |=
?
¬AXα iff s
1
S [(s,s
1
) R and (M, s
1
) |=
?
¬α],
21. (M,s) |=
?
¬EXα iff s
1
S [(s,s
1
) 6∈ R or (M, s
1
) |=
?
¬α],
22. (M,s) |=
?
¬AGα iff there exists a path π s
0
,s
1
,s
2
,...
with s s
0
, and for some state s
i
along π, we have
(M,s
i
) |=
?
¬α,
23. (M,s) |=
?
¬EGα iff for any paths π s
0
,s
1
,s
2
,...
with s s
0
, there exists a state s
i
along π such that
(M,s
i
) |=
?
¬α,
24. (M,s) |=
?
¬AFα iff there exists a path π s
0
,s
1
,s
2
,...
with s s
0
, and for any states s
i
along π, we have
(M,s
i
) |=
?
¬α,
25. (M,s) |=
?
¬EFα iff for any paths π s
0
,s
1
,s
2
,... with
s s
0
, and any states s
i
along π, we have (M, s
i
) |=
?
¬α,
26. (M,s) |=
?
¬A(αUβ) iff there exists a path π
s
0
,s
1
,s
2
,... with s s
0
, and for any states s
j
along π,
we have (M,s
j
) |=
?
¬β or 0 k < j (M, s
k
) |=
?
¬α,
27. (M,s) |=
?
¬E(αUβ) iff for any paths π s
0
,s
1
,s
2
,...
with s s
0
, and any states s
j
along π, we have
(M,s
j
) |=
?
¬β or 0 k < j (M, s
k
) |=
?
¬α,
28. (M,s) |=
?
¬A(αRβ) iff there exists a path π
s
0
,s
1
,s
2
,... with s s
0
, and for some state s
j
along π,
we have (M, s
j
) |=
?
¬β and 0 k < j (M,s
k
) |=
?
¬α,
29. (M,s) |=
?
¬E(αRβ) iff for any paths π s
0
,s
1
,s
2
,...
with s s
0
, there exists a state s
j
along π such that
(M,s
j
) |=
?
¬β and 0 k < j (M, s
k
) |=
?
¬α.
A formula α is called n-valid in CTL if (M, s) |=
?
α holds for any falsification-aware normal CTL-
model M := (S,S
0
,R,L
?
), any s S, and any
falsification-aware normal CTL-satisfaction relation
|=
?
on M.
Theorem 3.2. For any falsification-aware normal
CTL-model M := (S, S
0
,R,L
?
), any falsification-
aware normal CTL-satisfaction relation |=
?
on M,
any formula α, and any s S, we have: (M,s) |=
?
α
iff (M, s) 6|=
?
¬α.
Proof. By induction on α. We show some cases.
1. Case α p Φ: We obtain: (M,s) |=
?
p iff p
L
?
(s) iff ¬p 6∈ L
?
(s) iff (M,s) 6|=
?
¬p.
2. Case α βγ: We obtain: (M,s) |=
?
βγ
iff (M,s) 6|=
?
β or (M,s) |=
?
γ iff (M,s) 6|=
?
β
or (M,s) 6|=
?
¬γ (by induction hypothesis) iff
(M,s) 6|=
?
¬(βγ).
3. Case α ¬β: We obtain: (M,s) |=
?
¬β iff
(M,s) 6|=
?
β (by induction hypothesis with contra-
position) iff (M, s) 6|=
?
¬¬β.
4. Case α AXβ: We obtain: (M, s) |=
?
AXβ
iff s
1
S [(s, s
1
) 6∈ R or (M,s
1
) |=
?
β] iff
s
1
S [(s, s
1
) 6∈ R or (M,s
1
) 6|=
?
¬β] (by in-
duction hypothesis) iff not-[s
1
S [(s, s
1
) R
and (M,s
1
) |=
?
¬β]] iff not-[(M, s) |=
?
¬AXβ] iff
(M,s) 6|=
?
¬AXβ.
5. Case α AGβ: We obtain: (M,s) |=
?
AGβ iff
for any paths π s
0
,s
1
,s
2
,... with s s
0
, and
any states s
i
along π, we have (M,s
i
) |=
?
β iff
for any paths π s
0
,s
1
,s
2
,... with s s
0
, and
any states s
i
along π, we have (M, s
i
) 6|=
?
¬β (by
induction hypothesis) iff not-[there exists a path
π s
0
,s
1
,s
2
,... with s s
0
, and for some state s
i
along π, we have (M,s
i
) |=
?
¬β] iff not-[(M,s) |=
?
¬AGβ] iff (M, s) 6|=
?
¬AGβ.
6. Case α A(βUγ): We obtain: (M,s) |=
?
A(βUγ)
iff for any paths π s
0
,s
1
,s
2
,... with s s
0
, there
exists a state s
j
along π such that (M, s
j
) |=
?
γ
and 0 k < j (M,s
k
) |=
?
β iff for any paths π
s
0
,s
1
,s
2
,... with s s
0
, there exists a state s
j
along
π such that (M,s
j
) |=
?
γ and 0 k < j (M, s
k
) 6|=
?
¬β iff not-[there exists a path π s
0
,s
1
,s
2
,... with
s s
0
, and for any states s
j
along π, we have
(M,s
j
) |=
?
¬γ or 0 k < j (M, s
k
) |=
?
¬β] iff
not-[(M,s) |=
?
¬A(βUγ)] iff (M,s) 6|=
?
¬A(βUγ).
Remark 3.3. We make the following remarks.
1. Theorem 3.2 shows that the mapping condition
“for any p Φ and any s S, p L
?
(s) iff ¬p 6∈
L
?
(s) in Definition 3.1 can be extended to the
falsification-aware normal CTL-satisfaction rela-
tion for any formula α.
2. By the contraposition of the statement of Theorem
3.2, we can obtain the following standard Boolean
negation condition for falsification-aware nor-
mal CTL-satisfaction relation: (M,s) |=
?
¬α iff
(M,s) 6|=
?
α.
3. We use Theorem 3.2 (with the fact discussed just
above) for proving the equivalence between the
falsification-aware normal and normal semantics
for CTL. Indeed, Theorem 3.2 will be used for
proving the case α ¬β in Lemma 3.4.
Prior to prove the equivalence between the
falsification-aware normal and normal Kripke-style
semantics for CTL (i.e., the equivalence between the
n-validity and the validity), we need to show the fol-
lowing two lemmas.
Lemma 3.4. For any CTL-model M := (S,S
0
,R,L)
and any CTL-satisfaction relation |= on M, we can
ICAART 2022 - 14th International Conference on Agents and Artificial Intelligence
246
construct a falsification-aware normal CTL-model
N := (S, S
0
,R,L
?
) and a falsification-aware normal
CTL-satisfaction relation |=
?
on N such that for any
formula α and any s S, (M,s) |= α iff (N,s) |=
?
α.
Proof. Let M be a CTL-model (S, S
0
,R,L) and |=
be a CTL-satisfaction relation on M. Then, we de-
fine a falsification-aware normal CTL-model N :=
(S,S
0
,R,L
?
) such that for any s S and any p Φ,
1. p L
?
(s) iff p L(s),
2. ¬p L
?
(s) iff p 6∈ L(s).
Then, we can obtain the mapping condition p
L
?
(s) iff ¬p 6∈ L
?
(s), because we have: p L
?
(s)
iff p L(s) iff ¬p 6∈ L
?
(s).
We now prove this lemma by induction on α. We
show some cases.
1. Case α p Φ: We obtain: (M,s) |= p iff p
L(s) iff p L
?
(s) iff (N,s) |=
?
p.
2. Case α βγ: We obtain: (M, s) |= βγ
iff (M, s) 6|= β or (M,s) |= γ iff (N,s) 6|=
?
β or
(N,s) |=
?
γ (by induction hypothesis) iff (N,s) |=
?
βγ.
3. Case α ¬β: We obtain: (M,s) |= ¬β iff
(M,s) 6|= β iff (N,s) 6|=
?
β (by induction hypoth-
esis) iff (N,s) |=
?
¬β (by Theorem 3.2).
4. Case α AXβ: We obtain: (M,s) |= AXβ iff
s
1
S [(s,s
1
) 6∈ R or (M,s
1
) |= β] iff s
1
S [(s,s
1
) 6∈ R or (N,s
1
) |=
?
β (by induction hy-
pothesis) iff (N,s) |=
?
AXβ.
5. Case α AGβ: We obtain: (M, s) |= AGβ iff for
any paths π s
0
,s
1
,s
2
,... with s s
0
, and any
states s
i
along π, we have (M,s
i
) |= β iff for any
paths π s
0
,s
1
,s
2
,... with s s
0
, and any states
s
i
along π, we have (N,s
i
) |=
?
β (by induction hy-
pothesis) iff (N,s) |=
?
AGβ.
6. Case α A(βUγ): We obtain: (M,s) |= A(βUγ)
iff for any paths π s
0
,s
1
,s
2
,... with s s
0
, there
exists a state s
j
along π such that (M, s
j
) |= γ
and 0 k < j (M,s
k
) |= β iff for any paths π
s
0
,s
1
,s
2
,... with s s
0
, there exists a state s
j
along
π such that (N,s
j
) |=
?
γ and 0 k < j (N,s
k
) |=
?
β (by induction hypothesis) iff (N, s) |=
?
A(βUγ).
Lemma 3.5. For any falsification-aware normal
CTL-model M := (S,S
0
,R,L
?
) and any falsification-
aware normal CTL-satisfaction relation |=
?
on M, we
can construct a CTL-model N := (S,S
0
,R,L) and a
CTL-satisfaction relation |= on N such that for any
formula α and any s S, (M,s) |=
?
α iff (N,s) |= α.
Proof. Similar to the proof of Lemma 3.4.
Theorem 3.6 (Equivalence between n-validity and
validity in CTL). For any formula α, we have: α is
n-valid in CTL iff α is valid in CTL.
Proof. By Lemmas 3.4 and 3.5.
4 FALSIFICATION-AWARE DUAL
KRIPKE-STYLE SEMANTICS
FOR CTL
A falsification-aware dual Kripke-style semantics for
CTL is defined as follows.
Definition 4.1 (Falsification-aware dual Kripke-style
semantics for CTL). Let Φ be a non-empty set of
propositional variables and Φ
¬
be the set p | p
Φ} of negated propositional variables.
A structure (S,S
0
,R,L
+
,L
) is called a
falsification-aware dual CTL-model if
1. S is the set of states,
2. S
0
is a set of initial states and S
0
S,
3. R is a binary relation on S such that s S s
0
S [(s,s
0
) R],
4. L
+
and L
are mappings from S to the power set
of Φ such that for any p Φ and any s S, p
L
+
(s) iff p 6∈ L
(s).
A path in a falsification-aware dual CTL-model is
an infinite sequence of states, π = s
0
,s
1
,s
2
,... such
that i 0 [(s
i
,s
i+1
) R].
Falsification-aware dual CTL-satisfaction rela-
tions (M, s) |=
+
α and (M,s) |=
α for any formula
α, where M is a falsification-aware dual CTL-model
(S,S
0
,R,L
+
,L
) and s is a state in S, are defined in-
ductively by the following clauses:
1. for any p Φ, (M, s) |=
+
p iff p L
+
(s),
2. (M,s) |=
+
αβ iff (M,s) |=
+
α and (M,s) |=
+
β,
3. (M,s) |=
+
αβ iff (M,s) |=
+
α or (M,s) |=
+
β,
4. (M,s) |=
+
αβ iff (M,s) 6|=
+
α or (M,s) |=
+
β,
5. (M,s) |=
+
¬α iff (M,s) |=
α,
6. (M,s) |=
+
AXα iff s
1
S [(s,s
1
) 6∈ R or (M, s
1
) |=
+
α],
7. (M,s) |=
+
EXα iff s
1
S [(s,s
1
) R and (M,s
1
) |=
+
α],
8. (M,s) |=
+
AGα iff for any paths π s
0
,s
1
,s
2
,... with
s s
0
, and any states s
i
along π, we have (M, s
i
) |=
+
α,
9. (M,s) |=
+
EGα iff there exists a path π s
0
,s
1
,s
2
,...
with s s
0
, and for any states s
i
along π, we have
(M,s
i
) |=
+
α,
10. (M,s) |=
+
AFα iff for any paths π s
0
,s
1
,s
2
,...
with s s
0
, there exists a state s
i
along π such that
(M,s
i
) |=
+
α,
11. (M,s) |=
+
EFα iff there exists a path π s
0
,s
1
,s
2
,...
with s s
0
, and for some state s
i
along π, we have
(M,s
i
) |=
+
α,
Falsification-aware Semantics for CTL and Its Inconsistency-tolerant Subsystem: Towards Falsification-aware Model Checking
247
12. (M,s) |=
+
A(αUβ) iff for any paths π s
0
,s
1
,s
2
,...
with s s
0
, there exists a state s
j
along π such that
(M,s
j
) |=
+
β and 0 k < j (M,s
k
) |=
+
α,
13. (M,s) |=
+
E(αUβ) iff there exists a path π
s
0
,s
1
,s
2
,... with s s
0
, and for some state s
j
along π,
we have (M,s
j
) |=
+
β and 0 k < j (M,s
k
) |=
+
α,
14. (M,s) |=
+
A(αRβ) iff for any paths π s
0
,s
1
,s
2
,...
with s s
0
, and any states s
j
along π, we have
(M,s
j
) |=
+
β or 0 k < j (M,s
k
) |=
+
α,
15. (M,s) |=
+
E(αRβ) iff there exists a path π
s
0
,s
1
,s
2
,... with s s
0
, and for any states s
j
along π,
we have (M,s
j
) |=
+
β or 0 k < j (M,s
k
) |=
+
α,
16. for any p Φ, (M, s) |=
p iff p L
(s),
17. (M,s) |=
αβ iff (M,s) |=
α or (M,s) |=
β,
18. (M,s) |=
αβ iff (M,s) |=
α and (M,s) |=
β,
19. (M,s) |=
αβ iff (M,s) |=
+
α and (M,s) |=
β,
20. (M,s) |=
¬α iff (M,s) |=
+
α,
21. (M,s) |=
AXα iff s
1
S [(s,s
1
) R and (M,s
1
) |=
α],
22. (M,s) |=
EXα iff s
1
S [(s, s
1
) 6∈ R or (M,s
1
) |=
α],
23. (M,s) |=
AGα iff there exists a path π s
0
,s
1
,s
2
,...
with s s
0
, and for some state s
i
along π, we have
(M,s
i
) |=
α,
24. (M,s) |=
EGα iff for any paths π s
0
,s
1
,s
2
,...
with s s
0
, there exists a state s
i
along π such that
(M,s
i
) |=
α,
25. (M,s) |=
AFα iff there exists a path π s
0
,s
1
,s
2
,...
with s s
0
, and for any states s
i
along π, we have
(M,s
i
) |=
α,
26. (M,s) |=
EFα iff for any paths π s
0
,s
1
,s
2
,... with
s s
0
, and any states s
i
along π, we have (M, s
i
) |=
α,
27. (M,s) |=
A(αUβ) iff there exists a path π
s
0
,s
1
,s
2
,... with s s
0
, and for any states s
j
along π,
we have (M,s
j
) |=
β or 0 k < j (M,s
k
) |=
α,
28. (M,s) |=
E(αUβ) iff for any paths π s
0
,s
1
,s
2
,...
with s s
0
, and any states s
j
along π, we have
(M,s
j
) |=
β or 0 k < j (M,s
k
) |=
α,
29. (M,s) |=
A(αRβ) iff there exists a path π
s
0
,s
1
,s
2
,... with s s
0
, and for some state s
j
along π,
we have (M,s
j
) |=
β and 0 k < j (M,s
k
) |=
α,
30. (M,s) |=
E(αRβ) iff for any paths π s
0
,s
1
,s
2
,...
with s s
0
, there exists a state s
j
along π such that
(M,s
j
) |=
β and 0 k < j (M,s
k
) |=
α.
A formula α is called d-valid in CTL if (M,s) |=
+
α holds for any falsification-aware dual CTL-
model M := (S,S
0
,R,L
+
,L
), any s S, and any
falsification-aware dual CTL-satisfaction relations
|=
+
and |=
on M.
Lemma 4.2. For any CTL-model M := (S,S
0
,R,L)
and any CTL-satisfaction relation |= on M, we can
construct a falsification-aware dual CTL-model N :=
(S,S
0
,R,L
+
,L
) and falsification-aware dual CTL-
satisfaction relations |=
+
and |=
on N such that for
any formula α and any s S,
1. (M,s) |= α iff (N,s) |=
+
α,
2. (M,s) |= ¬α iff (N,s) |=
α.
Proof. Let M be a CTL-model (S, S
0
,R,L) and |=
be an CTL-satisfaction relation on M. Then, we
define a falsification-aware dual CTL-model N :=
(S,S
0
,R,L
+
,L
) such that for any s S and any
p Φ,
1. p L
+
(s) iff p L(s),
2. p L
(s) iff p 6∈ L(s).
Then, we can obtain the mapping condition p
L
+
(s) iff p 6∈ L
(s).
We now prove this lemma by simultaneous induc-
tion on α. We show some cases.
1. Case α p Φ: For 1, we obtain: (M,s) |= p iff
p L(s) iff p L
+
(s) iff (N,s) |=
+
p. For 2, we
obtain: (M,s) |= ¬p iff (M,s) 6|= p iff p 6∈ L(s) iff
p L
(s) iff (N,s) |=
p.
2. Case α βγ: For 1, we obtain: (M,s) |= βγ
iff (M,s) 6|= β or (M,s) |= γ iff (N,s) 6|=
+
β or
(N,s) |=
+
γ (by induction hypothesis for 1) iff
(N,s) |=
+
βγ. For 2, we obtain: (M, s) |=
¬(βγ) iff (M,s) 6|= βγ iff (M, s) |= β and
(M,s) 6|= γ iff (M, s) |= β and (M,s) |= ¬γ iff
(N,s) |=
+
β and (N,s) |=
γ (by induction hy-
potheses for 1 and 2) iff (N,s) |=
βγ.
3. Case α ¬β: For 1, we obtain: (M,s) |= ¬β iff
(N,s) |=
β (by induction hypothesis for 2) iff
(N,s) |=
+
¬β. For 2, we obtain: (M, s) |= ¬¬β
iff (M,s) |= β iff (N, s) |=
+
β (by induction hy-
pothesis for 1) iff (N,s) |=
¬β.
4. Case α AXβ: For 1, we obtain: (M,s) |= AXβ
iff s
1
S [(s, s
1
) 6∈ R or (M,s
1
) |= β] iff s
1
S [(s,s
1
) 6∈ R or (N,s
1
) |=
+
β (by induction hy-
pothesis for 1) iff (N,s) |=
+
AXβ. For 2, we ob-
tain: (M, s) |= ¬AXβ iff (M, s) 6|= AXβ iff not-
[(M,s) |= AXβ] iff not-[s
1
S [(s, s
1
) 6∈ R or
(M,s
1
) |= β],] iff s
1
S [(s, s
1
) R and (M,s
1
) 6|=
β] iff s
1
S [(s,s
1
) R and (M,s
1
) |= ¬β] iff
s
1
S [(s, s
1
) R and (N,s
1
) |=
β] (by induc-
tion hypothesis for 2) iff (N,s) |=
AXβ.
5. Case α AGβ: For 1, we obtain: (M,s) |= AGβ
iff for any paths π s
0
,s
1
,s
2
,... with s s
0
, and
any states s
i
along π, we have (M,s
i
) |= β iff for
any paths π s
0
,s
1
,s
2
,... with s s
0
, and any
states s
i
along π, we have (N,s
i
) |=
+
β (by induc-
tion hypothesis for 1) iff (N,s) |=
+
AGβ. For 2,
we obtain: (M, s) |= ¬AGβ iff (M, s) 6|= AGβ iff
there exists a path π s
0
,s
1
,s
2
,... with s s
0
, and
for some state s
i
along π, we have (M,s
i
) 6|= β iff
there exists a path π s
0
,s
1
,s
2
,... with s s
0
, and
for some state s
i
along π, we have (M, s
i
) |= ¬β iff
ICAART 2022 - 14th International Conference on Agents and Artificial Intelligence
248
there exists a path π s
0
,s
1
,s
2
,... with s s
0
, and
for some state s
i
along π, we have (N,s
i
) |=
β (by
induction hypothesis for 2) iff (N,s) |=
AGβ.
6. Case α A(βUγ): For 1, we obtain: (M,s) |=
A(βUγ) iff for any paths π s
0
,s
1
,s
2
,... with
s s
0
, there exists a state s
j
along π such that
(M,s
j
) |= γ and 0 k < j (M,s
k
) |= β iff for
any paths π s
0
,s
1
,s
2
,... with s s
0
, there ex-
ists a state s
j
along π such that (N,s
j
) |=
+
γ and
0 k < j (N,s
k
) |=
+
β (by induction hypoth-
esis) iff (N,s) |=
+
A(βUγ). For 2, we obtain:
(M,s) |= ¬A(βUγ) iff (M,s) 6|= A(βUγ) iff there
exists a path π s
0
,s
1
,s
2
,... with s s
0
, and
for any states s
j
along π, we have (M,s
j
) 6|= γ
or 0 k < j (M,s
k
) 6|= β iff there exists a path
π s
0
,s
1
,s
2
,... with s s
0
, and for any states
s
j
along π, we have (M,s
j
) |= ¬γ or 0 k < j
(M,s
k
) |= ¬β iff there exists a path π s
0
,s
1
,s
2
,...
with s s
0
, and for any states s
j
along π, we have
(N,s
j
) |=
γ or 0 k < j (N, s
k
) |=
β (by in-
duction hypothesis for 2) iff (N,s) |=
A(βUγ).
Lemma 4.3. For any falsification-aware dual CTL-
model M := (S,S
0
,R,L
+
,L
) and any falsification-
aware dual CTL-satisfaction relations |=
+
and |=
on
M, we can construct a CTL-model N := (S,S
0
,R,L)
and a CTL-satisfaction relation |= on N such that for
any formula α and any s S,
1. (M,s) |=
+
α iff (N,s) |= α,
2. (M,s) |=
α iff (N,s) |= ¬α.
Proof. Similar to the proof of Lemma 4.2.
Theorem 4.4 (Equivalence between d-validity and
validity in CTL). For any formula α, we have: α is
d-valid in CTL iff α is valid in CTL.
Proof. By Lemmas 4.2 and 4.3.
Remark 4.5. We can obtain the following fact by
Theorems 3.6 and 4.4. The following conditions are
equivalent: For any formula α,
1. α is valid in CTL,
2. α is n-valid in CTL,
3. α is d-valid in CTL.
5 FALSIFICATION-AWARE
KRIPKE-STYLE SEMANTICS
FOR ICTL
We now semantically define ICTL as a subsystem of
CTL. Formulas of ICTL are defined in the same way
as those for CTL.
Definition 5.1 (Falsification-aware normal Krip-
ke-style semantics for ICTL). The falsification-aware
normal Kripke-style semantics for ICTL is obtained
from that for CTL by deleting the mapping condi-
tion “for any p Φ and any s S, p L
?
(s) iff
¬p 6∈ L
?
(s)” in Definition 3.1. The notions and nota-
tions concerning this semantics for ICTL can be de-
fined in a similar way as those for CTL.
Remark 5.2. We make the following remarks.
1. We cannot semantically define ICTL using the
standard non-falsification-aware Kripke-style se-
mantics as presented in Definition 2.2.
2. ICTL is regarded as or closely related to the clas-
sical negation-free subsystems of PCTL (Kamide
and Kaneiwa, 2010; Kaneiwa and Kamide, 2011)
and pCTL (Kamide and Endo, 2018; Kamide and
Endo, 2019). Namely, ICTL is regarded as the
purely inconsistency-tolerant subsystem of PCTL
and pCTL.
3. As will be discussed in the following item 5, ICTL
is a proper subsystem of CTL, because the for-
mula of the form (p∧¬p)q, where p and q are
distinct propositional variables, is n-valid in CTL,
but not n-valid in ICTL.
4. The same theorem as Theorem 3.2 does not hold
for ICTL (i.e., the negation connective ¬ in ICTL
is not the classical negation connective).
5. In general, a satisfaction relation |= for a logic
is called paraconsistent with respect to a nega-
tion connective ¬ if the condition “α,β (M,s) 6|=
(α∧¬α)β holds. This condition reflects that
(α∧¬α)β is not valid in the underlying logic.
Based on this definition of paraconsistency, ICTL
is shown to be paraconsistent with respect to
¬, although CTL is not paraconsistent with re-
spect to ¬. The paraconsistency of ICTL with
respect to ¬ can be shown as follows. Assume
a falsification-aware normal ICTL-model M =
(S,S
0
,R,L
?
) such that p L
?
(s), ¬p L
?
(s)
and q / L
?
(s) for a pair of distinct propositional
variables p and q. Then, we have (M, s) 6|=
?
(p∧¬p)q.
6. ICTL is regarded as a four-valued logic. This fact
is shown as follows. For each s S and each
formula α, we can take one of the following four
cases:
(a) α is verified at s, i.e., (M, s) |=
?
α,
(b) α is falsified at s, i.e., (M, s) |=
?
¬α,
(c) α is both verified and falsified at s,
(d) α is neither verified nor falsified at s.
Falsification-aware Semantics for CTL and Its Inconsistency-tolerant Subsystem: Towards Falsification-aware Model Checking
249
7. We can show the fact that ICTL is embeddable
into the positive (i.e., ¬-less) fragment of CTL.
This fact can be proved in the same way as pre-
sented in (Kamide and Endo, 2018; Kamide and
Endo, 2019). Thus, the paraconsistent negation
connective in ICTL can be handled in the positive
fragment of CTL.
Next, we introduce a falsification-aware dual
Kripke-style semantics for ICTL.
Definition 5.3 (Falsification-aware dual Kripke-style
semantics for ICTL). By deleting the mapping con-
dition “for any p Φ and any s S, p L
+
(s)
iff p 6∈ L
(s) in Definition 4.1, we can obtain a
falsification-aware dual Kripke-style semantics for
ICTL.
Then, we can obtain the following theorem.
Theorem 5.4 (Equivalence between n-validity and d–
validity in ICTL). For any formula α, we have: α is
n-valid in ICTL iff α is d-valid in ICTL.
Proof. We can prove this theorem in a similar way as
the proof of Theorem 3.6. We have to prove the lem-
mas that are similar to Lemmas 4.2 and 4.3. But, to
prove these lemmas, we do not need to use Theorem
3.2 because ICTL has no classical negation.
6 CONCLUDING REMARKS
In this paper, we first introduced new falsification-
aware normal and dual Kripke-style semantics for
CTL. We then proved the equivalences among
the proposed falsification-aware Kripke-style seman-
tics and the standard Kripke-style semantics for
CTL. By using the proposed falsification-aware
Kripke-style semantics, we semantically defined the
new inconsistency-tolerant and many-valued tempo-
ral logic ICTL, which is obtained from the pro-
posed falsification-aware Kripke-style semantics for
CTL by deleting only one characteristic condition
on the labeling function of the semantics. Thus,
the proposed falsification-aware semantic framework
for CTL and ICTL is considered to be a unified
framework for combining and generalizing the stan-
dard, inconsistency-tolerant, and many-valued se-
mantic frameworks. The proposed framework can be
used for falsification-aware model checking, which
is roughly defined as that of combined and gener-
alized model checking based on falsification-aware
Kripke-style semantics. The notion of falsification-
aware model checking and an illustrative example for
falsification-aware model checking will be explained
and addressed in the last part of this section.
We remark that the proposed falsification-aware
semantic framework can be extended to other tem-
poral logics. For example, we can obtain the
falsification-aware normal and dual Kripke-style se-
mantics for linear-time temporal logic (LTL) (Pnueli,
1977) and its inconsistency-tolerant subsystem ILTL
in a similar manner as those for CTL and ICTL.
Then, we can prove the equivalences among the
falsification-aware Kripke-style semantics and the
standard Kripke-style semantics for LTL as well as
those between the falsification-aware normal and dual
Kripke-style semantics for ILTL. We can also obtain
the falsification-aware normal and dual Kripke-style
semantics for full computation tree logic, referred
to as CTL
(Emerson and Halpern, 1986; Emer-
son and Sistla, 1984), and its inconsistency-tolerant
subsystem ICTL
. We can also prove the equiv-
alences among the falsification-aware Kripke-style
semantics and the standard Kripke-style semantics
for CTL
as well as those between the falsification-
aware normal and dual Kripke-style semantics for
ICTL
. Thus, we can obtain a falsification-aware
model checking framework based on these standard
and inconsistency-tolerant temporal logics.
Furthermore, we can also obtain the falsification-
aware normal and dual Kripke-style semantics for
probabilistic CTL, referred to as pCTL, probabilis-
tic CTL
, referred to as pCTL
(Aziz et al., 1995;
Bianco and de Alfaro, 1995), and their inconsistency-
tolerant subsystems IpCTL and IpCTL
. We can also
prove the equivalences among the falsification-aware
Kripke-style semantics and the standard Kripke-style
semantics for pCTL and pCTL
, respectively, as well
as those between the falsification-aware normal and
dual Kripke-style semantics for IpCTL and IpCTL
,
respectively. In addition to these results, we can
also obtain the falsification-aware normal and dual
Kripke-style semantics for hierarchical (or sequential)
CTL, referred to as sCTL, hierarchical (or sequential)
LTL, referred to as sLTL (Kamide and Yano, 2017;
Kamide, 2018), and their inconsistency-tolerant sub-
systems IsCTL and IsLTL, respectively. Then, we can
prove the equivalences among the falsification-aware
Kripke-style semantics and the standard Kripke-style
semantics for sCTL and sLTL, respectively, as well as
those between the falsification-aware normal and dual
Kripke-style semantics for IsCTL and IsLTL, respec-
tively. Thus, we can also obtain a falsification-aware
model checking framework based on these extended
temporal logics.
Finally, we illustrate a small verification exam-
ple based on falsification-aware model checking. We
consider the disease diagnosis model shown in Figure
1 for familial adenomatous polyposis (FAP) (Wehbi,
ICAART 2022 - 14th International Conference on Agents and Artificial Intelligence
250
Figure 1: A diagnosis model for familial adenomatous polyposis.
2019). FAP is an inherited disorder characterized by
an autosomal dominant inherited condition in which
numerous adenomatous polyps form mainly in the ep-
ithelium of the large intestine. Although these polyps
start out benign, malignant transformation into colon
cancer occurs when they are left untreated. Three
variants are known to exist: FAP, attenuated FAP
(AFAP), and MUTYH-associated polyposis (MAP).
Furthermore, the first variant of FAP is classified as
dense- or non-dense-type FAP. For these variants, the
resulting colonic polyps and cancers were initially
confined to the colon wall. Detection and removal be-
fore metastasis outside the colon can greatly reduce
and eliminate the spread of cancer in many cases.
For the model presented in Figure 1, we can verify
the following statements:
1. “Is there a state in which a person is both healthy
and unhealthy (i.e., not yet ill), has less than 100
adenomas, and had a medical examination with
genetic diagnosis?”
2. “Is there a state in which a person has a family
medical history, more than 100 adenomas, had a
medical examination with clinical diagnosis, and
has MAP?”
Although the first statement is true because there is a
case in which the person has multiple benign adeno-
mas, the second statement is false because this case
does not imply MAP but implies AFAP. These state-
ments are formally expressed as follows:
1. EF(healthy∧¬healthy(adenoma<100)
geneticDiagnosis)
2. EF(hasMedicalHistoryFamily(adenoma>100)
clinicalDiagnosisMAP)
where the negation connective ¬ in the first formula is
regarded as the inconsistency-tolerant negation con-
nective in ICTL. In this example, we can simulta-
neously verify and falsify these formulas using the
falsification-aware dual CTL- or ICTL-satisfaction
relations |=
+
and |=
in the falsification-aware dual
Kripke-style semantics. Furthermore, we can for-
mally consider a falsification-aware model checking
problem as follows. Suppose that M is a falsification-
aware dual CTL- or ICTL-model (S,S
0
,R,L
+
,L
)
and that |=
+
and |=
are falsification-aware dual
CTL- or ICTL-satisfaction relations on M. Then, the
falsification-aware model checking problem for CTL
or ICTL is defined as follows. For any formula α,
find the verification set {s S | M, s |=
+
α} and fal-
sification set {s S | M,s |=
α}. These sets can
be simultaneously found. Finally, we remark that
extended falsification-aware pCTL-, IpCTL-, sCTL-
, and IsCTL-model checking frameworks, which are
based on the aforementioned extended CTLs, are
more suitable for the verification of the model pre-
sented in Figure 1 because probabilistic and hierar-
chical specifications are also required to verify this
model.
ACKNOWLEDGEMENTS
This research was supported by JSPS KAKENHI
Grant Numbers JP18K11171 and JP16KK0007.
REFERENCES
Almukdad, A. and Nelson, D. (1984). Constructible falsity
and inexact predicates. Journal of Symbolic Logic,
49:231–233.
Falsification-aware Semantics for CTL and Its Inconsistency-tolerant Subsystem: Towards Falsification-aware Model Checking
251
Aziz, A., Singhal, V., and Balarin, F. (1995). It usually
works: The temporal logic of stochastic systems. In
Proceedings of the 7th Int. Conf. on Computer Aided
Verification (CAV 1995), Lecture Notes in Computer
Science 939, pages 155–165.
Bianco, A. and de Alfaro, L. (1995). Model checking of
probabilistic and nondeterministic systems. In Pro-
ceedings of the 15th Conf. on Foundations of Soft-
ware Technology and Theoretical Computer Science
(FSTTCS 1995), Lecture Notes in Computer Science
1026, pages 499–513.
Chen, D. and Wu, J. (2006). Reasoning about inconsistent
concurrent systems: A non-classical temporal logic.
In Lecture Notes in Computer Science, volume 3831,
pages 207–217.
Clarke, E. and Emerson, E. (1981). Design and synthesis of
synchronization skeletons using branching time tem-
poral logic. In Lecture Notes in Computer Science,
volume 131, pages 52–71.
Clarke, E., Grumberg, O., Jha, S., Lu, Y., and Veith, H.
(2003). Counterexample-guided abstraction refine-
ment for symbolic model checking. Journal of the
ACM, 50 (5):752–794.
Clarke, E., Henzinger, T., Veith, H., and Bloem, R. (2018).
Handbook of Model Checking. Springer.
da Costa, N., Beziau, J., and Bueno, O. (1995). Aspects of
paraconsistent logic. Bulletin of the IGPL, 3 (4):597–
614.
Easterbrook, S. and Chechik, M. (2001). A framework for
multi-valued reasoning over inconsistent viewpoints.
In Proceedings of the 23rd International Conference
on Software Engineering (ICSE 2001), pages 411–
420.
Emerson, E. and Halpern, J. (1986). ‘sometimes’ and ‘not
never’ revisited: on branching versus linear time tem-
poral logic. Journal of the ACM, 33 (1):151–178.
Emerson, E. and Sistla, A. (1984). Deciding full branching
time logic. Information and Control, 61:175–201.
Gurfinkel, A., Wei, O., and Chechik, M. (2006). Yasm:
A software model-checker for verification and refuta-
tion. In Proceedings of the 18th International Con-
ference on Computer Aided Verification (CAV 2006),
Lecture Notes in Computer Science, volume 4144,
pages 170–174.
Horn, L. and Wansing, H. (2017). Negation. The Stan-
ford Encyclopedia of Philosophy (Spring 2017 Edi-
tion), Edward N. Zalta (editor), Last modified on Jan-
uary 2017.
Kamide, N. (2006). Extended full computation-tree logics
for paraconsistent model checking. Logic and Logical
Philosophy, 15 (3):251–276.
Kamide, N. (2015). Inconsistency-tolerant temporal rea-
soning with hierarchical information. Information Sci-
ences, 320:140–155.
Kamide, N. (2018). Logical foundations of hierarchical
model checking. Data Technologies and Applications,
52 (4):539–563.
Kamide, N. (2021). Falsification-aware semantics and se-
quent calculi for classical logic. Journal of Philosoph-
ical Logic, Online first article.
Kamide, N. and Bernal J.P.A. (2019). Towards loca-
tive inconsistency-tolerant hierarchical probabilistic
ctl model checking: Survey and future work. In
Proceedings of the 11th International Conference on
Agents and Artificial Intelligence (ICAART 2019),
volume 2, pages 869–878.
Kamide, N. and Endo, K. (2018). Logics and transla-
tions for inconsistency-tolerant model checking. In
Proceedings of the 10th International Conference on
Agents and Artificial Intelligence (ICAART 2018),
volume 2, pages 191–200.
Kamide, N. and Endo, K. (2019). Foundations of
inconsistency-tolerant model checking: Logics, trans-
lations, and examples. In Agents and Artificial In-
telligence, the 10th International Conference ICAART
2018 Revised Selected Papers, Lecture Notes in Arti-
ficial Intelligence, volume 11352, pages 1–31.
Kamide, N. and Kaneiwa, K. (2010). Paraconsistent nega-
tion and classical negation in computation tree logic.
In Proceedings of the 2nd International Conference
on Agents and Artificial Intelligence (ICAART 2010),
volume 1, pages 464–469.
Kamide, N. and Koizumi, D. (2015). Combining paracon-
sistency and probability in ctl. Proceedings of the 7th
International Conference on Agents and Artificial In-
telligence (ICAART 2015), 2:285–293.
Kamide, N. and Koizumi, D. (2016). Method for combin-
ing paraconsistency and probability in temporal rea-
soning. Journal of Advanced Computational Intelli-
gence and Intelligent Informatics, 20:813–827.
Kamide, N. and Wansing, H. (2011). A paraconsistent
linear-time temporal logic. Fundamenta Informaticae,
106 (1):1–23.
Kamide, N. and Yamamoto, Y. (2021). Inconsistency-
tolerant hierarchical probabilistic computation tree
logic and its application to model checking. Proceed-
ings of the 13th International Conference on Agents
and Artificial Intelligence (ICAART 2021), 2:490–
499.
Kamide, N. and Yano, R. (2017). Logics and translations
for hierarchical model checking. Proceedings of the
21st International Conference on Knowledge-Based
and Intelligent Information and Engineering Systems
(KES2017), Procedia Computer Science, 112:31–40.
Kaneiwa, K. and Kamide, N. (2011). Paraconsistent com-
putation tree logic. New Generation Computing, 29
(4):391–408.
Nelson, D. (1949). Constructible falsity. Journal of Sym-
bolic Logic, 14:16–26.
Pnueli, A. (1977). The temporal logic of programs. In Pro-
ceedings of the 18th IEEE Symposium on Foundations
of Computer Science, pages 46–57.
Priest, G. and Tanaka, K. (2009). Paraconsistent logic. Web
site of the Stanford encyclopedia of philosophy (2009
Edition), Edward N. Zalta (editor).
Wansing, H. (1993). The logic of information structures.
In Lecture Notes in Computer Science, volume 681,
pages 1–163.
Wehbi, M. (2019). Familial adenomatous polyposis. Web
site of eMedicine Gastroenterology.
ICAART 2022 - 14th International Conference on Agents and Artificial Intelligence
252