Professional Documents
Culture Documents
(Artigo) A New Algorithmic Decision For Categorical Syllogisms Via Carroll's Diagrams (2019)
(Artigo) A New Algorithmic Decision For Categorical Syllogisms Via Carroll's Diagrams (2019)
(Artigo) A New Algorithmic Decision For Categorical Syllogisms Via Carroll's Diagrams (2019)
https://doi.org/10.1007/s00500-019-04598-9
Abstract
In this paper, we propose a new effective algorithm for the categorical syllogisms by using a calculus system Syllogistic
Logic with Carroll Diagrams, which determines a formal approach to logical reasoning with diagrams, for representations
of the fundamental Aristotelian categorical syllogisms. We show that this logical reasoning is closed under the syllogistic
criterion of inference. Therefore, the calculus system is implemented to let the formalism which comprises synchronically
bilateral and trilateral diagrammatical appearance and naive algorithmic nature. And also, there is no need specific knowledge
or exclusive ability to understand this decision procedure as well as to use it in an algorithmic system. Consequently, the
empirical contributions of this paper are to design a polynomial-time algorithm at the first time in the literature to conduce
to researchers getting into the act in different areas of science used categorical syllogisms such as artificial intelligence,
engineering, computer science and etc.
123
N. Kircali Gursoy et al.
123
A new algorithmic decision for categorical syllogisms via Carroll’s diagrams
AAA E AE I AI AE E A AI A AO AE O S exists
E AE AE E AI I I AI E AO E AO S exists
AI I EIO O AO EIO A AI E AO M exists
EIO AO O EIO EIO M exists
A AI P exists
In the categorical syllogistic system, there are 64 different We describe a homomorphic mapping between the categor-
syllogistic forms for each figure. These are called moods. ical syllogistic propositions and Carroll’s diagrams for the
Therefore, the categorical syllogistic system is composed of calculus system SLCD.
256 possible syllogisms. Only 24 of them are valid in this Let X and Y be two terms and let X and Y be com-
system. And they divided into two groups of 15 and 9. plements of X and Y , respectively. For two terms, Carroll
The syllogisms in the first group are valid unconditionally divides the square into four cells, by this means he gets the
which are given in Table 3. so-called bilateral diagram, as given in Table 5.
The syllogisms in the second group called strengthened Each of these four cells can have three possibilities when
syllogism are valid conditionally or valid existential import, we explain the relations between two terms. They can be 0 or
which is an explicit supposition of being of some terms, are 1 or blank. In this method, 0 means that there is no element
given in Table 4. intersection cell of two elements, 1 means that it is not empty
and blank cell means that we don’t have any idea about the
content of the cell, it could be 0 or 1.
3 Representation of categorical syllogisms As in the above method, let X , Y and M be three terms and
via Carroll’s diagrams and a calculus X , Y and M be complements of X , Y and M, respectively.
system SLCD For examining all relations between three terms, he added
one more square in the middle of the bilateral diagram which
In this section, first of all we state propositions and their is called the trilateral diagram, as in Table 6.
relations with each other by the help of Carroll bilateral and Each cell in a trilateral diagram is marked with a 0, if there
trilateral diagrams, respectively. Then, we explain how all the is no element and is marked with an I if it is not empty and
possible diagrams of these bilateral diagrams are obtained another using of I, it could be on the line where the two cells
and give a mapping that assigns a value to each possible is intersection, this means that at least one of these cells is
123
N. Kircali Gursoy et al.
Table 5 Relation of two terms X X Table 7 The paradigm for the representation of syllogistic arguments
Y XY XY by using sets
Y XY XY LOGIC DIAGRAMS SETS
T ranslate
PREMISES Propositions −−−−−−−→ Sets
↓
T ranslate
CONCLUSIONS Propositions ←−−−−−−− Sets
Table 6 Relations of three
terms
j =
r val 2(4−i) n i , 1 ≤ j ≤ k,
We give the set-theoretical representation of syllogistic argu- i=1
ments by means of bilateral diagrams. To build such a model,
where k is the number of all possible forms.
we draw from Carroll’s diagrammatic method. We give a def-
inition of a map which corresponds each bilateral diagram
to a set. Eventually, our main purpose is to construct a com- Definition 2 Let R set be the set of the values which cor-
plete bridge between sets and categorical syllogisms such as respond to all possible forms of any bilateral diagram;
Table 7. that is R set = {r val
j : 1 ≤ j ≤ k, k is the number
123
A new algorithmic decision for categorical syllogisms via Carroll’s diagrams
of all possible forms}. The set of all these R set ’s is denoted Table 13 X intersection Y has X X
by RSet .
at least one element R(I) = Y
Y 1
Corollary 1 We obtain the set representations of all categor-
ical propositions as follows:
Table 14 X intersection Y has X X
– All X are Y It means that X intersection with Y cell is the at least one element R(O) = Y 1
empty set. We can illustrate this statement as Table 11. Y
From Table 11, we obtain all possible forms as the same
method in Example 1. By the help of Definition 1, the
set representation of “All X are Y” corresponds to the Table 15 The possible forms of bilateral diagrams
set = {0, 1, 2, 3, 8, 9, 10, 11}.
R(A)
P P S S
– No X are Y There is no element in the intersection cell of
pi = M 1 0 and pj = M 0 1
X and Y as Table 12. M 0 0 M 0 0
By Example 1, we have all possible forms of “No X are
Y”. Then, we obtain R(E) set = {0, 2, 4, 6, 8, 10, 12, 14}.
Example 2 Let pi and p j be two possible forms of the bilat- we obtain that rk = 2 corresponds to pk that is a possible
eral diagrams of major and minor premises, respectively. We conclusion.
take the possible forms of bilateral diagrams as Table 15.
We input the data on the trilateral diagram as in Table 16. Let rival and r val
j be the numbers corresponding to possible
By using the elimination method, we obtain the relation forms of bilateral diagrams that have a common term. Then
between S and P as Table 17. ri = 8 corresponds to possible we can get the relation between two other terms by using this
form pi , and r j = 4 corresponds to possible form p j , then method.
123
N. Kircali Gursoy et al.
3.3 Constructing operation of possible conclusion Table 19 The conclusion of the major and the minor premises
table P P S S P P
M M = S
M M S
After these above examples, we generalize all possible con-
clusions by using only a formula.
123
A new algorithmic decision for categorical syllogisms via Carroll’s diagrams
n
t
Framework of the whole algorithmic structure is as fol-
set
R(k) R(l)
set
:= rkval
j
∗ rlval
i lows:
j=1 i=1
set and R set . Step 1 Conditional selection from the conditions set
is the conclusion of the premises R(k) (l) {U nconditional, S_exists, M_exists, P_exists}.
Theorem 3 (Senturk and Oner 2019) A syllogism is valid if Step 2 Figure selection from the figures set {Figur eI ,
and only if it is provable in SLCD. Figur eI I , Figur eI I I , Figur eI V }.
Step 3 Selection of the Major proposition from propositions
Remark 1 For conditional valid forms, we need an addition set {A, E, I , O}.
rule which is “Some X are X ”. We can use the above theorem Step 4 Selection of the Minor proposition from propositions
by taking into consideration this rule. set {A, E, I , O}.
Step 5 Determination of the valid syllogisms.
Remark 2 Let SLCD be noted calculus system. If the rule
“Some X are X when X exists” (i.e., I X X ) is added to In the above framework, firstly, a condition is selected
SLCD, then the calculus system SLCD is denoted by SLCD† . from four conditional cases (U nconditional, S_exists,
Definition 6 (Senturk and Oner 2019) Let R(k) be the bilat- M_exists, P_exists) in Step 1. In step 2, a figure is deter-
eral diagram presentation of the premise. The transposition mined from four figures (Figur eI , Figur eI I , Figur eI I I ,
of a premise is the symmetric positions with respect to the Figur eI V ). Then, by examining all proposition states (A,
main diagonal. It is shown by Trans(R(k) ). E, I , O) of major and minor premises, conclusion sets are
created with using operation table of possible conclusions
Trans : Rset → Rset , (Table 18). Lastly, the four conclusion sets of syllogism (from
the relation between S and P) are compared with the conclu-
set
R(k) → T rans(R(k)
set
) = {rkval
T , . . . , r k T }.
set
1 n sion set obtained from the premises and valid syllogisms are
determined.
Theorem 4 (Senturk and Oner 2019) Let R(k) set = {r val , . . . ,
k1 Global variables used to all algorithms are below:
rkset
n
} and R set = {r val , . . . , r val } be two sets to correspond
(l) l1 lt Conc[ ][ ]: two dimensional array, set of all possible bilateral
to the Major and the Minor premises values sets and R(s) set = diagrams
{rs1 , . . . , rsm } be set to correspond to the constant set values
val set Const_set: constant set for each condition S exists, M exists
which means “Some S are S”, “Some M are M” and “Some and P exists
P are P”. Then † : Rset Maj × RMin → RCon
set set
4.1 The syllogism algorithm
R(l)
set † set
R(k)
⎧ n t m This is the main algorithm. In this algorithm (Algorithm 1),
⎪ val
⎪
⎪ (rk j ∗ (rsvar ∗ rlvar
T )), If S exists AT P() and Decision() subalgorithms are run for each state
⎪
⎪ h
⎪
⎪ j=1 i=1 h=1 i (Unconditional, S exists, M exists and P exists) and for each
⎨ n t m figure (FigureI, FigureII, FigureIII and FigureIV ) as defined
:= (rkval ∗ (rlvar ∗ rsvar )), If M exists
⎪
⎪
j i h in Table 2. This algorithm sends the related figure as a param-
⎪
⎪
j=1 i=1 h=1
n t
⎪
⎪
m eter to the AT P() subalgorithm and also, it sends the related
⎪
⎩ ((rsvar ∗ rkval
T ) ∗ rli ),
var If P exists
j=1 i=1 h=1
h
j state and figure as parameters to the Decision() subalgo-
rithm.
is the conclusion of the premises R(k)set and R set under the
(l)
conditions S exists, M exists or P exists. 4.2 The adjusting of terms positions (ATP)
subalgorithm
Theorem 5 (Senturk and Oner 2019) A strengthened syllo-
gism is valid if and only if it is provable in SLCD† . This subalgorithm (Algorithm 2) determines the positions of
the subject term, middle term and predicate term for major
and minor premises with respect to the figure parameter in
4 An algorithmic decision for categorical input as Table 2.
syllogisms in SLCD
4.3 The decision subalgorithm
In this part of this work, we give an algorithm to decide
whether a categorical syllogism is valid or not in the calculus This subalgorithm (Algorithm 3) determines major and
system SLCD or SLCD† for the first time in the literature. minor sets for each proposition (A, E, I and O) of major and
123
N. Kircali Gursoy et al.
– If the state is “S exists” and the premise type is “minor”, 9 if pr emise_t ype = “ma jor then
then new temp set is determined by taking the transpose 10 if cond = “P E xists then
11 N T emp_set =
of temp set and it returns the result of the subalgorithm
transpose the diagram o f T emp_set
Syllogistic_Mapping() which gets inputs as the con- 12 return Syllogistic_Mapping(Const_set, N T emp_set)
stant set and the new temp set, respectively.
13 r etur n T emp_set
– If the state is “M exists” and the premise type is
“minor”, then it returns the result of the subalgorithm
123
A new algorithmic decision for categorical syllogisms via Carroll’s diagrams
4.5 An example of the algorithm in SLCD Using above notation, the time complexity of the main
algorithm is T (m, n, k) = 2mnk 3 + 6mnk 2 + 5mnk + 8mn.
In this example, we choose “All M are P” and “All S are Furthermore, if taking m = n = k, the syllogism algorithm
M” for major and minor premises, respectively under the asymptotically has O(n 5 ) running time. Hence, the presented
unconditional statement. algorithm runs in polynomial time.
Initially, the algorithm determines major and minor
premises sets.
All M are P All S are M 5 Conclusion
0-0-0-0 0-0-0-0
1-0-0-0 1-0-0-0 In this paper, we present a new effective algorithm for the
0-1-0-0 0-0-1-0 categorical syllogisms by using a calculus system SLCD for
1-1-0-0 1-0-1-0 the first time in literature. In accordance with this purpose,
0-0-0-1 0-0-0-1 we explain categorical syllogisms by the help of Carroll’s
1-0-0-1 1-0-0-1 diagrams and we find unconditionally valid syllogisms and
0-1-0-1 0-0-1-1 conditionally valid syllogisms via this algorithmic approach.
1-1-0-1 1-0-1-1 As a result, our aim in this paper is to design a polynomial-
Major Premise Set: Minor Premise Set: time algorithm to contribute to researchers getting into the
{0, 8, 4, 12, 1, 9, 5, 13} {0, 8, 2, 10, 1, 9, 3, 11} act in different areas of science used categorical syllogisms
such as artificial intelligence, engineering, computer science
Then, the algorithm obtains possible conclusions from
and also mathematics.
major and minor premises sets by using the ∗-operation in
Table 18 and refines the repetitive values in conclusion set. Acknowledgements The authors would like to thank all the reviewers
Note that, some operation conclusions are not listed because for their constructive comments. This study is supported by Ege Univer-
they don’t have any possible conclusion. sity Scientific Research Projects Directorate with the Project Number
21267.
0∗0=0 8∗8=8 4∗8=4 12 ∗ 8 = 12
1∗2=4 1∗1=1 1∗3=5 9 ∗ 10 = 12 Compliance with ethical standards
9∗9=9 9 ∗ 11 = 13 5 ∗ 10 = 4 5∗9=5
5 ∗ 11 = 5 13 ∗ 10 = 12 13 ∗ 9 = 13 13 ∗ 11 = 13 Conflict of interest The authors declare that they have no conflict of
interests.
Conclusion set: {0, 8, 4, 12, 4, 1, 5, 12, 9, 13, 4, 5, 5, 12, Human and animal participants This article does not contain any stud-
13, 13}. ies with human participants performed by any of the authors.
After omitting the duplicated elements:
Conclusion set: {0, 1, 4, 5, 8, 9, 12, 13}.
Finally, we compare with the four conclusion sets of syllo- References
gism (from the relation between S and P) and the conclusion
set obtained from the premises. It is determined that AAA is Abeles FF (2005) Lewis Carroll’s formal logic. Hist Philos Logic
valid. 26(1):33–46
Abeles FF (2007) Lewis Carroll’s visual logic. Hist Philos Logic
28(1):1–17
(1) AAA-VALID Barnes J (2014) The complete works of Aristotle: the revised Oxford
(2) AAE- translation. Princeton University Press, Princeton
(3) AAI- Carroll L (1958) Symbolic logic and the game of logic. Courier Corpo-
(4) AAO- ration, North Chelmsford
Castro-Manzano JM (2015) Słupecki’s rule for diagrammatic reason-
ing. Studia Metodologiczne 35:79–96
4.6 The complexity of the main algorithm Castro-Manzano JM (2017) Re(dis)covering Leibniz’s diagrammatic
logic. Tópicos, Revista de Filosofía 52:89–116
Deng W, Zhao H, Zou L, Li G, Yang X, Wu D (2017) A novel collab-
In this section, we briefly give a complexity analysis of the
orative optimization algorithm in solving complex optimization
main algorithm. The entries used in the main algorithm and problems. Soft Comput 21(15):4387–4398
the subalgorithms as below: Glasgow J, Narayanan NH, Chandrasekeran B (1995) Diagrammatic
reasoning: cognitive and computational perspectives. AAAI Press,
Palo Alto
m: the number of conditions,
Kulinkovich AE (1979) Algorithmization of reasoning in solving
n: the number of figures, geological problems. In: Proceedings of the methodology of geo-
k: the number of propositions. graphical sciences, pp 145–161
123
N. Kircali Gursoy et al.
Kumova BI, Cakir H (2010) Algorithmic decision of syllogisms. In: Rocha J, Meseguer J (2007) A rewriting decision procedure for
International conference on industrial, engineering and other appli- Dijkstra–Scholten’s syllogistic logic with complements. Revista
cations of applied intelligent systems. Springer, Heidelberg Colombiana de Comp 8(2):101–130
Łukasiewicz J (1957) Aristotle’s syllogistic from the standpoint of mod- Senturk I, Oner T (2019) An algebraic analysis of categorical syllogisms
ern formal logic, 2nd edn. Oxford University Press, Oxford by using Carroll’s diagrams. Filomat 33(2):367–383
Mayer RE (1992) Thinking, problem solving, and cognition. WH Free- Senturk I, Oner T, Nuriyev U (2016) Completeness of categorical syl-
man, New York logisms by means of diagrammatic method. In: Association for
Mayer RE (2009) Multimedia learning, 2nd edn. Cambridge University symbolic logic European summer meeting, logic colloquium 2016,
Press, New York Leeds, England
Moktefi A (2008) Lewis Carroll’s logic. In: Gabbay DM, Woods J (eds) Skliar O, Monge RE, Gapper S (2015) Using inclusion diagrams as an
The handbook of the history of logic, vol. 4: British logic in the alternative to Venn diagrams to determine the validity of categor-
nineteenth-century. Elsevier, Amsterdam, pp 457–505 ical syllogisms. arXiv:1509.00926
Moktefi A (2015) A history of logic diagrams, visual reasoning with Turunen E (2014) An algebraic study of Peterson’s intermediate syllo-
diagrams. Springer, Berlin, pp 55–71 gisms. Soft Comput 18(12):2431–2444
Moktefi A (2018) Carroll diagrams: design and manipulation. In: Chap- Zadeh LA (1985) Syllogistic reasoning as a basis for combination of
man P, Stapleton G, Moktefi A, Perez-Kriz S, Bellucci F (eds) evidence in expert systems. In: 9th international joint conference
Diagrammatic representation and inference. Diagrams, lecture on artificial intelligencem pp 417–419
notes in computer science, vol 10871. Springer, Cham, pp 32–35 Zhao H, Zheng J, Xu J, Deng W (2019a) Fault diagnosis method based
Moktefi A, Shin SJ (2012) A history of logic diagrams, handbook of the on principal component analysis and broad learning system. IEEE
history of logic, vol 11. North-Holland, Amsterdam, pp 611–682 Access 7:99263–99272
Moktefi A (2017) Diagrams as scientific instruments. In: Benedek A, Zhao H, Liu H, Xu J, Deng W (2019b) Performance prediction using
Veszelszki A (eds) Visual, virtual, veridical. Visual learning, vol high-order differential mathematical morphology gradient spec-
7. Peter Lang Verlag, Frankfurt/M., pp 81–89 trum entropy and extreme learning machine. IEEE Trans Instrum
Nakatsu RT (2014) Using Venn diagrams to perform logic reasoning: an Meas. https://doi.org/10.1109/TIM.2019.2948414
algorithm for automating the syllogistic reasoning of categorical
statements. Int J Intell Syst 29(1):84–103
Niittymäki J, Turunen E (2003) Traffic signal control on similarity logic Publisher’s Note Springer Nature remains neutral with regard to juris-
reasoning. Fuzzy Sets Syst 133:109–131 dictional claims in published maps and institutional affiliations.
Pratt-Hartmann I, Moss LS (2009) On the computational complexity of
the numerically definite syllogistic and related logics. Rev Symb
Logic 2:647–683
123