=Paper=
{{Paper
|id=Vol-2672/paper_1
|storemode=property
|title=How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs
|pdfUrl=https://ceur-ws.org/Vol-2672/paper_1.pdf
|volume=Vol-2672
|authors=Markus Brenneis,Maike Behrendt,Stefan Harmeling,Martin Mauve
|dblpUrl=https://dblp.org/rec/conf/comma/BrenneisBHM20
}}
==How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs==
https://ceur-ws.org/Vol-2672/paper_1.pdf
2 Brenneis et al. / How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs
How Much Do I Argue Like You?
Towards a Metric on Weighted
Argumentation Graphs
Markus BRENNEIS a,1 , Maike BEHRENDT a Stefan HARMELING a and
Martin MAUVE a
a Department of Computer Science, University of Düsseldorf, Germany
Abstract. When exchanging arguments with other people, it is interesting to know
who of the others has the most similar opinion to oneself. In this paper, we sug-
gest using weighted argumentation graphs that can model the relative importance
of arguments and certainty of statements. We present a pseudometric to calculate
the distance between two weighted argumentation graphs, which is useful for ap-
plications like recommender systems, consensus building, and finding representa-
tives. We propose a list of desiderata which should be fulfilled by a metric for those
applications and prove that our pseudometric fulfills these desiderata.
Keywords. argumentation graphs, online argumentation, metric
1. Introduction
In real-world discussions, people exchange arguments on a dedicated issue, such as im-
proving the course of study [11], the distribution of funds [7], or which party to vote for
at the next general election. In all those cases, participants discuss positions like “there
should be a universal basic income” or “special math courses should be introduced”,
state their pro and contra arguments and attack other people’s arguments.
Each individual participant of an argumentation has a personal view on the argu-
ments and their relative importance: Users can decide for themselves which arguments
they consider more convincing, thus which arguments they agree to and how much they
agree with a statement. They may consider some positions more important than others.
Based on those individual views, there are useful applications for measuring the
similarity or distance between two users: Clustering can be used to find representatives
for a group of people with similar argumentation behavior or for finding a consensus.
Another application is opinion polling, where one wants to find out why two persons or
organizations come to different conclusions. What is more, collaborative filtering, which
needs some definition of distance between users, can be used for pre-filtering arguments
in applications like Kialo2 .
Copyright c 2020 for this paper by its authors. Use permitted under Creative Commons License Attribution
4.0 International (CC BY 4.0).
1 Corresponding Author: Markus Brenneis, Heinrich-Heine-Universität, Universitätsstraße 1, 40225
Düsseldorf, Germany; E-mail: markus.brenneis@uni-duesseldorf.de.
2 https://kialo.com/
Brenneis et al. / How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs 3
In this paper, we propose solutions to the two main challenges to achieving the goal
of comparing the argumentations of two users: We define weighted argumentation graphs
which are a suitable representation of argumentation covering the mentioned aspects,
including importance of arguments and agreement with statements. Secondly, we suggest
a pseudometric for calculating the distance between two weighted argumentation graphs,
which considers the specific structure of argumentation graphs (e.g., opinions deeper in
a graph are less important). We contribute a list of useful desiderata for a metric which
compares argumentations, and prove that our pseudometric fulfills those properties.
In the following chapter, we present our definition of weighted argumentation
graphs. The third chapter introduces our pseudometric and desiderta for a useful metric.
Finally, we discuss some limitations of our pseudometric and take a look at related work.
2. Definition of Weighted Argumentation Graphs
To be able to determine the similarity of real-world argumentations, there has to be a
suitable representation of them. This representation should be able to capture all aspects
mentioned in the introduction, and it should be as simple as possible. Therefore, the
following definition is based on the IBIS model [12], which has been successfully tested
with users without background in argumentation theory using our D-BAS system [11].
For the application purposes described in the introduction, the model should be able
to represent the known opinions and arguments of a person as close as possible. Thus, we
use statements, not arguments, as atomic elements in our definition, which then can be
composed to arguments which can support or attack another statement. Note, though, that
this definition can be translated to classical abstract argumentation frameworks based on
Dung’s definition [5], e.g. using DABASCO [14].
Let S be the (finite) set of all statements, with the special statement I ∈ S. The set
A ⊆ S \ {I} × S is a set of arguments. For an argument a = (s1 , s2 ) ∈ A, s1 is called
premise, s2 conclusion of a. Let s ∈ S, then a→s := {(t, u) ∈ A | u = s} is the set of
arguments with conclusion s.
Note that I is excluded to be a premise since it is the issue in the IBIS model. We
refer to I as “personal well-being”, which allows us to interpret an edge like (b, I) in
Figure 1 as “My personal well-being improves, because more wind power plants will be
built.” The premises of arguments with conclusion I are called positions. Positions are
actionable items like “A wall between Austria and Germany should be built,” and play
an important role in real-world argumentation, e.g. decision-making problems [7].
Definition 1 (Argumentation Graph). An argumentation graph is a directed, weakly con-
nected graph G = (S, A), A ⊆ S \ {I} × S, where the statements are nodes and the argu-
ments edges, and there is exactly one I ∈ S which has no outgoing edges.
Note that this model does not include different relations for attack and support, as
known from bipolar argumentation frameworks. Whether an argument is supportive are
not, is up to a person’s interpretation of the natural language representation of the argu-
ments. The purpose of the model is solely to capture the hierarchy of statement, which
we later need for our metric; bipolarity would add unnecessary complexity in this paper.
Every person can have a personal view with personal attitudes on a common argu-
mentation graph G, as depicted in G0 in Figure 1. To get an intuition for our next defini-
4 Brenneis et al. / How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs
G I G0 on G 0 I
1 2
3 3
a A wall between Austria and Germany should be built.
a b −0.2 a 0.4 b a1 The wall stops illegal immigration of cows.
0.2 0.8 a11 Cows can come via Switzerland.
a2 The wall is expensive.
a1 a2 b1 0.3 a1 0.5 a2
b More wind power plants should be built.
b1 Wind power is a renewable energy source.
a11
Figure 1. Example for an argumentation graph G and a weighted argumentation graph G0 on G with positions
a and b and concrete examples for each statement. Edges with weight 0 and nodes with rating 0 are not drawn.
Statement ratings are next to nodes, values for relative argument importance next to edges.
tion, let us have a look at what we can deduce about Alice’s attitudes from her graph G0 :
Alice strongly accepts position b (rating .4) and is slightly against a (rating −.2). She
accepts the statement a2 more than statement a1 (rating .5 > rating .3). The counterargu-
ment (a2 , a) “No wall should be built, because a wall is expensive.” is far more important
for her than the argument (a1 , a) (relative importance .8 > .2).
Furthermore, it makes sense to sort the positions: She considers building more wind
power plants (b) more important than building a wall (a, relative importance 23 > 31 ).
So when comparing her attitudes with someone else’s attitudes, she would consider a
contrary opinion on b more severe than a different opinion on a. For ordinary statements,
which are not positions, having an importance does not make sense: One cannot say
that “The wall stops illegal immigration of cows” is twice as important as “Cows can
come via Switzerland” (important regarding what?); one can only say that the arguments
regarding building a wall which are built by those statements are of differing importance.
We will use real numbers to represent those weights and ratings.
Definition 2 (Weighted Argumentation Graph). Let G = (S, A) be an argumentation
graph. A weighted argumentation graph G0 on G is a quadruple (S, A, r, w) with func-
tions r and w. r : S → [−0.5, 0.5] assigns an agreement score (rating) to every statement,
where negative values mean disagreement, 0 no opinion/don’t care, and positive values
agreement. w : A → [0, 1] assigns an importance weight to each argument. The value in-
dicates the importance of that argument relative to other arguments with the same con-
clusion. The value 0 means that the argument is not used (i.e. has no relevance), and 1
means that the argument is the only relevant argument for the conclusion. The following
conditions must hold:
∀s ∈ S ∑ w(a) ∈ {0, 1} (1)
a∈A→s
r(I) = 0 (2)
Formula 1 means that the sum of weights of arguments with the same conclusion is
1 if there is an argument with positive weight (cf. (a1 , a) and (a2 , a) in Figure 1); the sum
is 0 iff no argument for a common conclusion has a weight (cf. w(b1 , b) = 0 in Figure 1).
This assures that w represents relative, not absolute importance. To simplify notation, we
write w(·, ·) instead of w((·, ·)). If the underlying argumentation graph G happens to be
a directed tree, we call G0 a weighted argumentation tree.
w and r can be represented as matrix or vector, respectively, where undefined values
are set to the default value 0. For the example in Figure 1, one gets:
Brenneis et al. / How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs 5
0 0 00000 0
0.33 0 0 0 0 0 0 −0.2
0.67 0 0 0 0 0 0 0.4
0 0.2 0 0 0 0 0 ,
w= r=
0.3
(3)
0 0.8 0 0 0 0 0 0.5
0 0 0 0 0 0 0 0
0 0 00000 0
The entry in row i column j of w is the weight of the argument with premise i and
conclusion j. The first column and first row must refer to I as premise or conclusion,
respectively. Because of Formula 1, the column sum is always 0 or 1.
If we draw or talk about a weighted argumentation graph, “non-existing” edges a
are edges with w(a) = 0 (argument with no importance), and “non-existing” nodes s are
nodes with r(s) = 0 (neutral statement). An example is G0 shown in Figure 1.
The importance of a position p is represented as weight of the “argument” (p, I)
leading to the “personal well-being” I. An application could obtain weights and ratings
from a user, for example, by asking them to mark statements which are considered more
important, or sorting arguments by relevance, which we do in our deliberate system [4].
3. Proposal of a Pseudometric for Weighted Argumentation Graphs
We now propose a pseudometric for calculating a distance between two weighted argu-
mentation graphs, and prove several properties we expect of a function which compares
two argumentations. The goal of the metric is to indicated how close the opinions and
used arguments of two persons are, considering graph structure and individual assess-
ments of importance; we do not want to compare argumentations on abstract levels like
consistency, number of arguments used, or if other person’s arguments are countered.
3.1. The Pseudometric
We define a distance measure of two weighted argumentation graphs G1 = (S, A, r1 , w1 ),
G2 = (S, A, r2 , w2 ) on G = (S, A) as:
∞
dG (G1 , G2 ) = (1 − α) ∑ α i kwi1 [:,1] r1 − wi2 [:,1] r2 k1 (4)
i=1
where α ∈ (0, 1) determines the influence of opinions deeper in the graph: A lower α
emphasizes opinions on statements r(s), a higher value the similarity of the argumen-
tation underneath a statement s. w[:,1] denotes the first column of the weight matrix w,
α i is the i-th power of the scalar α, wi the i-th power of the square matrix w, and
the Hadamard (entrywise) product. The i-th summand calculates the contribution of the
paths with length i ending at I. We drop the index G if the underlying argumentation
graph is clear from the context.
The intuition behind this distance measure becomes clearer when rephrasing it for
the special case of argumentation trees (which have no cycles or re-used statements, thus
unique paths from each statement to I). In case of argumentation trees T1 and T2 on T ,
the distance dG is equivalent to:
6 Brenneis et al. / How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs
dT (T1 , T2 ) = (1 − α) ∑ α depth(s) r1 (s) ∏ w1 (a) − r2 (s) ∏ w2 (a) (5)
s∈S a∈ρI→s a∈ρI→s
ρs1 →s2 is the sequence of all arguments (edges) on the path from s1 to s2 ∈ S (where
s2 is deeper in the tree). depth(s) = |ρI→s | is the length of the path from I to s, i.e. the
number of arguments; depth(I) = 0.
The terms in the absolute value measure the similarity of the opinions of a statement
s as difference of their ratings, scaled with the product of the “importances” of the ar-
guments leading to s. Hence, statements which are deeper in the argumentation tree get
a smaller weight, and the overall relevance of an argumentation branch is limited to its
importance.
To see how the calculation works and that the results match intuition, let us calculate
the distance between the graphs G0 (Figure 1), T2 , and T3 (Figure 2) for α = 0.5. We can
expect that T2 is closer to T3 than to G0 , because the opinions on the statements a and b
match and only the weights are different. The results confirm the expectation:
1 2
d(T2 , G0 ) =(1 − α) α 1 0.5 · 0.6 − (−0.2) · + α 1 0.5 · 0.4 − 0.4 ·
3 3
(6)
1 1
+ α 0 · 0 · 0.6 − 0.3 · 0.2 · + α 2 0 · 0 · 0.6 − 0.5 · 0.8 ·
2
= 0.194
3 3
d(T2 , T3 ) =(1 − α)(α 1 |0.5 · 0.6 − 0.5 · 0.3| + α 1 |0.5 · 0.4 − 0.5 · 0.7| = 0.075 (7)
Note that the value of dG and dT is in [0, 1). If d is the depth of T , the maximum
value of dT (T1 , T2 ) is α(1 − α d ).3 The maximum value of dG is lim (1 − α) ∑ni=1 α i = α.
n→∞
Theorem 1. Let G be an argumentation graph. dG is a pseudometric, i.e. has the follow-
ing properties for all weighted argumentation graphs G1 , G2 , G3 on G:
(i) dG (G1 , G1 ) = 0
(ii) dG (G1 , G2 ) = dG (G2 , G1 ) (symmetry)
(iii) dG (G1 , G3 ) ≤ dG (G1 , G2 ) + dG (G2 , G3 ) (triangle inequality)
Proof. dG converges: ∑i α i is geometric series, which converges for α ∈ (0, 1). The value
of the L1 norm cannot be greater than 1 because for each column sum σ of the wi , we
always have 0 ≤ σ ≤ 1.
(i) holds because the same values are subtracted in the L1 norm.
(ii) is given since the L1 norm is symmetric.
As each summand fulfills the triangle inequality, dG also fulfills (iii).
However, dG is not a metric, because dG (G1 , G2 ) can be 0 even if G1 is not equal to
G2 : Consider G1 where all statements are agreed to and every argument weight is 0, and
G2 where all statements have a rating of 0. Because weights and ratings are multiplied,
the distance is 0 though the weighted argumentation graphs are different.
3 Remember that r (root(T )) = r (root(T )) because root(T ) = I, and r(I) := 0.
1 2
Brenneis et al. / How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs 7
3.2. Desiderata for a Metric for Weighted Argumentation Trees
Our pseudometric is only one of many possible metrics for the applications described.
We present a list of intuitive desiderata which, as we think, should be fulfilled by any
metric comparing two person’s argumentations, and thus should be considered when
constructing alternative metric proposals. It is, however, hard to capture intuitive prop-
erties in graphs which can contain circular references and re-used statements. Therefore,
we focus on argumentation trees in this section. Field experiments like [11] have also
shown that users seldom create cycles or re-use statements in different branches in real
discussions.
After each desideratum, we prove that our pseudometric (Formula 5) fulfills it for
weighted argumentation trees. We consider those properties important in many real-
world application domains of a metric, albeit not everywhere, as pointed out in Section 5.
For each desideratum, we indicate why we think it is intuitive. Most desiderata are
followed by a visual example making the choice of variable names clearer. Note that
each tree in the examples is considered to be part of a bigger weighted argumentation
tree, i.e. not all existing nodes and edges are drawn, and irrelevant statement ratings and
argument weights are left out.
Desideratum 1 (Proportionally bigger overlap is better). Consider trees T1 , T2 , T3 , where
T2 is like T1 , but uses one additional argument for a statement s, and T3 is like T2 , but
uses one additional argument for s. Although T2 and T1 differ in only one argument, and
T3 and T2 differ in only one argument, we expect d(T1 , T2 ) > d(T2 , T3 ) because T2 and T3
have a greater overlap regarding the used arguments.
More formally: For every statement s in a tree T which only has leaves s1 , . . . , sn
(n > 2) as premises for s with r(s1 ) = · · · = r(sn ) 6= 0 and w(s1 , s) = · · · = w(sn , s) and
∀a ∈ ρI→s : w(a) 6= 0, consider the trees Tk , n ≥ k > 0, which only contain s1 , . . . , sk .
Then, given k1 < k2 < k3 , we want to have kk21 < kk23 =⇒ d(Tk1 , Tk2 ) > d(Tk2 , Tk3 ), i.e. if the
relative overlap of the number of arguments used is greater, the distance is smaller. Like-
wise, we demand kk21 > kk23 =⇒ d(Tk1 , Tk2 ) < d(Tk2 , Tk3 ), and kk12 = kk32 =⇒ d(Tk1 , Tk2 ) =
d(Tk2 , Tk3 ).
We require ∀a ∈ ρI→s : w(a) 6= 0 (i.e. no argument (edge) with weight 0 along the
path from I to s), because if a user gives an argument a weight of 0, they say the premise
is not related to the conclusion, thus not related to the topic of the discussion. This means
that the user actually does not care about the opinions underneath that argument, which
may be treated as if no opinion has been given.
In the following example, we have k1 = 1, k2 = 2, and k3 = 3, thus 12 < 23 =⇒
dT (T1 , T2 ) > dT (T2 , T3 ):
T1 s T2 s T3 s
1 1 1 1 1 1
k1 k2 k2 k3 k3 k3
r s1 r s1 r s2 r s1 r s2 r s3
Proof. Only the argument weights for (si , s) (namely k11 , k12 , k13 , respectively, as used
in (8) = (9)) are different and contribute to the sum, all other summands are zero. The
common weight products and common values for r are summarized as wr for readability
and are factored out. For (10) > (11), remember that kk12 < kk32 .
8 Brenneis et al. / How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs
0
dT (Tk1 , Tk2 ) = wr (1 − α) ∑ α depth(s ) |wk1 (s0 , s) − wk2 (s0 , s)| (8)
s0 ∈{s1 ,...,sn }
1
depth(s0 ) 1 1
= wr (1 − α) ∑ α k1 · − + (k2 − k1 ) · (9)
s0 ∈{s1 ,...,sn }
k1 k2 k2
depth(s0 ) k1
= wr (1 − α) ∑ α 2 − 2 (10)
s0 ∈{s1 ,...,sn }
k2
depth(s0 ) k2
> wr (1 − α) ∑ α 2 − 2 = dT (Tk2 , Tk3 ) (11)
s0 ∈{s ,...,s }
k3
1 n
The other cases are proven by replacing “>” with “<” or “=”, respectively.
Desideratum 2 (Contrary opinion is worse than no opinion). Consider trees T1 , T2 , T3 ,
where all trees are identical, but T1 has no opinion on a statement s, T2 a positive opinion
on s and T3 a negative opinion on s. As we definitely know that T2 and T3 disagree on s,
we want to have d(T2 , T3 ) > d(T1 , T2 ).
Formally: For any statement s in a tree T with ∀a ∈ ρI→s : w(a) 6= 0, let T + be like
T but with r+ (s) = q > 0, T − like T with r− (s) = p < 0, and T 0 like T with r0 (s) = 0.
Then d(T + , T − ) > d(T + , T 0 ).
T0 0 s T+ q s T− p s
Proof. The only positive summand is the summand for s (which has rating 0, q or p for
T 0 , T + , and T − , respectively). The argument weights and the α term are common to all
summands, can be factored out, and are summarized as wα .
dT (T + , T − ) = wα · |q − p| = wα · (q + p) > wα · |q − 0| = dT (T + , T 0 ) (12)
Desideratum 3 (Deviation in deeper parts has less influence than deviation in higher
parts). Consider the trees T1 , T2 , T3 , where T1 has an argument (sA , s) with no chil-
dren and ∀a ∈ ρI→s : w(a) 6= 0, T2 is constructed from T1 by adding a new statement
sB and argument (sB , s) with w2 (sB , s) = w2 (sA , s) = w1 (s2A ,s) , and T3 is constructed
from T1 by adding a new statement sA1 and argument (sA1 , sA ) with w3 (sA1 , sA ) = 1. If
r(sA ) = r(sB ) = r(sA1 ) 6= 0, then we want to have d(T1 , T2 ) > d(T1 , T3 ), because argu-
ments deeper in the tree should have a smaller influence since we consider them less
important for the overall opinion.
We require r(sA ) = r(sB ) = r(sA1 ) 6= 0 because adding a statement with “don’t care”
opinion should not actually change the distance. We want equality because this desider-
atum should cover only differences in the depth of the statements, not their rating.
T1 s T2 s T3 s
w w
w 2 2 w
r sA r sA r sB r sA
1
r sA1
Brenneis et al. / How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs 9
Proof. The only differences are the summands including s, sA and sA1 . Let r := r(sA ) =
r(sB ) = r(sA1 ) 6= 0 and w := w1 (sA , s). As before, we summarize common values for
weights and α as wα .
w w
dT (T1 , T2 ) = wα · rw − r + 0−r = wα · |rw| (13)
2 2
> wα · α|rw| = wα · (α · |r · 0 · w − r · 1 · w|) = dT (T1 , T3 ) (14)
Desideratum 4 (Influence of deeper parts depends on weights in higher parts). Consider
trees T1 , T2 , T3 , where all trees are identical, have statements sA and sB with conclusion s,
and w(sA , s) > w(sB , s), but T1 has an additional argument for sA and T3 has an additional
argument for sB . Although the difference in both cases is only one argument, we expect
d(T1 , T2 ) > d(T3 , T2 ) because (sA , s) has a larger weight.
Formally: Let T2 be a weighted argumentation tree with arguments (sA , s), (sB , s) and
w(sA , s) > w(sB , s), no premises for sA and sB and ∀a ∈ ρI→s : w(a) 6= 0. T1 is constructed
from T2 by adding (sA1 , sA ) and T3 from T2 by adding (sB1 , sB ), each with a weight of 1
and r1 (sA1 ) = r3 (sB1 ) 6= 0. Then dT (T1 , T2 ) > dT (T3 , T2 ).
T1 s T2 s T3 s
w(sA , s) w(sB , s) w(sA , s) w(sB , s) w(sA , s) w(sB , s)
sA sB sA sB sA sB
1 1
r sA1 r sB1
Proof. Let r := r1 (sA1 ) = r3 (sB1 ) 6= 0. Only the summand which includes sA1 or sB1 ,
respectively, contributes a value greater than 0.
dT (T1 , T2 ) = wα · |w(sA , s) · r − 0| > wα · |0 − w(sB , s) · r| = dT (T2 , T3 ) (15)
Desideratum 5 (Weights of arguments have influence even if they are the only differ-
ence). Consider trees T1 , T2 , where all trees are identical and have the arguments (sA , s)
and (sB , s), but the weights are different: w1 (sA , s) 6= w2 (sA , s) and w1 (sB , s) 6= w2 (sB , s).
We want to have d(T1 , T2 ) > 0 if there exists a statement s0 below sA (or sA itself) with
r1 (s0 ) = r2 (s0 ) 6= 0 and ∀a ∈ ρI→s0 : w(a) 6= 0.
We demand r1 (s0 ) = r2 (s0 ) 6= 0 because it makes sense if weights leading only to
statements which are rated as “don’t care” are ignored.
In this example, we have sA = s0 :
T1 s T2 s
w1 (sA , s) w1 (sB , s) w2 (sA , s) w2 (sB , s)
r sA sB r sA sB
10 Brenneis et al. / How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs
Proof. It is enough to show that there is at least one summand greater than 0.
|w1 (sA , s) − w2 (sA , s)| > 0 (16)
0
=⇒ α depth(s ) (1 − α)|w1 (sA , s) − w2 (sA , s)| > 0 (17)
depth(s0 )
=⇒ α (1 − α)r1 (s0 ) ∏ w1 (sA0 )|w1 (sA , s) − w2 (sA , s)| > 0 (18)
sA0 ∈ρs0 →I \(sA ,s)
0
=⇒ α depth(s ) (1 − α) r1 (s0 ) ∏ w1 (sA0 ) − r2 (s0 ) ∏ w2 (sA0 ) > 0 (19)
sA0 ∈ρs0 →I sA0 ∈ρs0 →I
For (18) =⇒ (19), remember that all weights in ρs0 →I \ (sA , s) and all ratings are
the same for T1 and T2 , e.g. r1 (s0 ) = r2 (s0 ).
Desideratum 6 (Symmetry regarding negation of opinion). Let T1 , T2 be any weighted
argumentation trees, and T3 , T4 , respectively, the same trees, but the opinion for each
statement is negated, i.e. r3 (s) = −r1 (s) and r4 (s) = −r2 (s) for all s ∈ S. We expect that
a metric is symmetric regarding negation, i.e. d(T1 , T2 ) = d(T3 , T4 ).
Proof. This holds because |r1 (s) − r2 (s)| = |(−r3 (s)) − (−r4 (s))| = |r3 (s) − r4 (s)|.
Desideratum 7 (Trade-off between argument weights and agreement). Consider trees
T1 , T2 , T3 which are nearly identical and have leaf statements sA and sB with com-
mon conclusion s and ∀a ∈ ρI→s : w(a) 6= 0. We have r1 (sA ) = r2 (sA ) = 0.5, r3 (sA ) =
−0.5, r1 (sB ) = r2 (sB ) = r3 (sB ) = 0, and w1 (sA , s) > w2 (sA , s). Furthermore, w2 (sB , s) =
w1 (sB , s) + w1 (sA , s) − w2 (sA , s), w3 (sB , s) = w1 (sB , s) + w1 (sA , s) − w3 (sA , s), i.e. sB is
neutral and “collects” remaining weight such that the sum is 1.
If w1 (sA , s) − w2 (sA , s) > w2 (sA , s) + w3 (sA , s), although T1 and T2 have the same
opinion on sA , we want to have d(T1 , T2 ) > d(T2 , T3 ), because both T2 and T3 do not
care much about their (different) opinions on sA . Likewise, if w1 (sA , s) − w2 (sA , s) <
w2 (sA , s) + w3 (sA , s), we expect d(T1 , T2 ) < d(T2 , T3 ) because the weights w1 and w2 are
closer to each other and give a greater weight for opposing opinions on sA .
The following example trees depict the first case with concrete weight values. Be-
cause the different opinions on statement sA are underneath an argument edge with small
weight, we want to have d(T1 , T2 ) > d(T2 , T3 ).
T1 0 s T2 0 s T3 0 s
0.9 0.1 0.1 0.9 0.2 0.8
0.5 sA 0 sB 0.5 sA 0 sB −0.5 sA 0 sB
Proof. We proof the first case. For (20) = (21), remember that w1 (sA , s) > w2 (sA , s).
d(T1 , T2 ) = wα · |0.5 · w1 (sA , s) − 0.5 · w2 (sA , s)| (20)
= wα · 0.5 · (w1 (sA , s) − w2 (sA , s)) (21)
> wα · 0.5 · (w2 (sA , s) + w3 (sA , s)) (22)
= wα · |0.5 · w2 (sA , s) − (−0.5) · w3 (sA , s)| = d(T2 , T3 ) (23)
Brenneis et al. / How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs 11
The other case follows by replacing “>” with “<”.
Desideratum 8 (Trade-off between statement ratings and agreement). Consider trees
T1 , T2 , T3 which are nearly identical and have a statement s and ∀a ∈ ρI→s : w(a) 6= 0. We
have r1 (s) > r2 (s) > 0 > r3 (s) such that |r1 (s) − r2 (s)| > |r2 (s) − r3 (s)|. Although T1 and
T2 have the same positive opinion on s, we want to have d(T1 , T2 ) > d(T2 , T3 ), because
both T2 and T3 have a weak opinion on s. Likewise, if |r1 (s) − r2 (s)| < |r2 (s) − r3 (s)|, we
expect d(T1 , T2 ) < d(T2 , T3 ) because the ratings r1 (s) and r2 (s) are closer to each other
than r2 (s) and r3 (s).
The first case, d(T1 , T2 ) > d(T2 , T3 ), is shown in the following example:
T1 0.4 s T2 0.1 s T3 −0.1 s
Proof. Only the summand for s contributes to the distance, all other summands are 0.
Remember that all weights on the path to s are positive.
d(T1 , T2 ) = wα |r1 (s) − r2 (s)| > wα |r2 (s) − r3 (s)| = d(T2 , T3 ) (24)
The other case follows by replacing “>” with “<”.
Desideratum 9 (Weights limit the influence of a path). Consider graphs T1 , T2 which
are nearly identical, have an argument (sA , s) with w = w1 (sA , s) = w2 (sA , s) and only
the ratings and weights below (and including) sA may differ in any way. No matter how
those values are chosen, we want to have d(T1 , T2 ) ≤ w, i.e. the maximum influence of
the differences below sA is limited by the weight of sA to its conclusion.
Proof. It is enough to consider paths which include a, since the summands for all other
parts are 0. Let Sa be the set of all statements which have an argument leading to a,
0
including a itself. We abbreviate α depth(s ) (1 − α) as α(s0 ).
dT (T1 , T2 ) = ∑ α(s0 ) r1 (s0 ) ∏ w1 (sA0 ) − r2 (s0 ) ∏ w2 (sA0 ) (25)
s0 ∈Sa sA0 ∈ρS→I sA0 ∈ρS→I
= w ∑ α(s0 ) r1 (s0 ) ∏ w1 (sA0 ) − r2 (s0 ) ∏ w2 (sA0 )
s0 ∈Sa sA0 ∈ρS→I \(sA ,s) sA0 ∈ρS→I \(sA ,s)
(26)
As the factor after w is in [0, 1], we get dT (T1 , T2 ) ≤ w.
4. Limitations
Although the proposed pseudometric fulfills several intuitive desiderata, there are some
limitations which we will discuss in the following.
If a weight of an argument is 0, the proposed pseudometric ignores all weights which
are underneath this argument. When comparing how similar people argue for or against
12 Brenneis et al. / How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs
T1 0 I T2 0 I T3 0 I T30 0 I
0.1 0.9 0.6 0.4 0.3 0.7 0.15 0.35 0.5
0.5 a 0.5 b 0.5 a 0.5 b 0.5 a 0.5 b 0.5 a 0.5 b 0.5 c
Figure 2. Possible unexpected change in order if an unrelated opinion is added:
0.05 = d(T1 , T3 ) < d(T2 , T3 ) = 0.075, but 0.1375 = d(T1 , T30 ) > d(T2 , T30 ) = 0.125 with α = 0.5
the top-level positions, this is okay, but if also the way how arguments which are not
supported are attacked should influence the distance, the metric has to be extended.
Moreover, ordering can be changed by adding an unrelated opinion. Consider trees
T1 , T2 , T3 with d(T1 , T3 ) < d(T2 , T3 ). At first sight, it might seem unexpected that this
order can be changed to d(T1 , T30 ) > d(T2 , T30 ) by adding a new position c to T3 and
keeping the relative weights of the other positions. Depending on the application context,
e.g. a voting advice application (VAA), this might be unwanted. An example is depicted
in Figure 2. This is due to the normalization of the argument weights.
Although even end-user friendly systems like D-BAS support for undercuts, i.e. ar-
guments that have an argument as conclusion, undercuts are currently not explicitly mod-
eled in our model. This is no big problem, because in many applications, for instance, in
a VAA, arguments can be preselected such that no undercut attack is necessary (since the
arguments make sense), or rephrased such that the premise is attacked. For example, con-
sider the argument “We should build more nuclear power plants because cats are cute”
and the attack “Cats and nuclear power plants are unrelated”. Though this is technically
an undercut, a user interface may present this as an attack on “Cats are cute”.
5. Related Work
Calculating the distance between argumentation graphs to compare how similar the atti-
tudes of two agents are has already been used in other systems. The Carneades opinion
formation and polling tool presented in [9] is able to compare one’s argumentation with
the argumentation of other entities like organizations. This comparison is simply done
by counting the number of statements where the agreement/disagreement is the same.
This approach is much simpler than our proposal, but uses neither weights nor ratings
and violates i.a., Desideratum 3.
The mobile application described in [1] also bases on IBIS and extends it with an
agreement value for each argument in the argumentation tree. This information is used,
for opinion prediction using collaborative filtering. In contrast to our work, the idea of
relative argument importance in combination with statement rating is not present.
Another application of calculating the similarity of weighted trees is match-making
of agents, which are represented by weighted trees. In [3], a recursive similarity measure
for this application is proposed. Its parameter N serves a similar purpose to our parameter
α. They also give examples which are similar to our desiderata, e.g. Example 4 is like
our Desideratum 2. Nodes, however, do not have a weight, and some desiderata are not
fulfilled; for instance, Desideratum 5 is explicitly not demanded in their Example 2.
There are already other definitions of weighted argumentation graphs based upon
Dung’s definition of argumentation frameworks, but many lack the differentiation be-
tween argument relevance and statement ratings (e.g. [2,8,10,6]), which we think is im-
Brenneis et al. / How Much Do I Argue Like You? Towards a Metric on Weighted Argumentation Graphs 13
portant since argument weights limit how much a branch of an argumentation is rele-
vant, whereas statement ratings are only relevant for the single statement. Furthermore,
in most cases, there is a global assignment of values in the graph and no user-specific
views, which is why a strength of 0 for attacks would be meaningless in the model pre-
sented in [13]. Note that most related work in this field is concerned about evaluating
consistency or calculating extensions, whereas our main goal is comparing the attitudes
of agents, not caring about whether an agents’ attitude is logically consistent or not.
6. Conclusion and Future Work
In this paper we proposed a pseudometric to calculate the distance between two argu-
mentation graphs representing the attitudes of different persons, and several desiderata
which should be considered when proposing other metrics for the same purpose, and
are fulfilled by our pseudometric. Possible next steps include developing other sensible
metrics and comparing them regarding theoretical properties and practicality.
In future work, we want to check if the desiderata are not only intuitive for experts in
argumentation theory, but are following the intuition of untrained humans. We also want
to test the metric in a VAA to compare the argumentation of voters with those of parties.
Thereby we see whether the results of the metric are accepted in an application context.
References
[1] Althuniyan, N., Sirrianni, J.W., Rahman, M.M., Liu, X.F.: Design of mobile service of intelligent large-
scale cyber argumentation for analysis and prediction of collective opinions. In: International Confer-
ence on AI and Mobile Services. pp. 135–149. Springer (2019)
[2] Amgoud, L., Ben-Naim, J., Doder, D., Vesic, S.: Acceptability semantics for weighted argumentation
frameworks. In: IJCAI. vol. 2017, pp. 56–62 (2017)
[3] Bhavsar, V.C., Boley, H., Yang, L.: A weighted-tree similarity algorithm for multi-agent systems in
e-business environments. Computational Intelligence 20(4), 584–602 (2004)
[4] Brenneis, M., Mauve, M.: deliberate – online argumentation with collaborative filtering. In: COMMA
(2020), to appear
[5] Dung, P.M.: On the acceptability of arguments and its fundamental role in nonmonotonic reasoning,
logic programming and n-person games. Artificial Intelligence 77(2), 321–357 (1995)
[6] Dunne, P.E., Hunter, A., McBurney, P., Parsons, S., Wooldridge, M.: Weighted argument systems: Basic
definitions, algorithms, and complexity results. Artificial Intelligence 175(2), 457–486 (2011)
[7] Ebbinghaus, B.: Decision making with argumentation graphs. arXiv preprint arXiv:1908.03357 (2019)
[8] Gordon, T.F., Walton, D.: Formalizing balancing arguments. In: COMMA. pp. 327–338 (2016)
[9] Gordon, T.F.: Structured consultation with argument graphs. From Knowledge Representation to Argu-
mentation in AI. A Festschrift in Honour of Trevor Bench-Capon on the Occasion of his 60th Birthday
pp. 115–133 (2013)
[10] Hunter, A.: A probabilistic approach to modelling uncertain logical arguments. International Journal of
Approximate Reasoning 54(1), 47–81 (2013)
[11] Krauthoff, T., Meter, C., Mauve, M.: Dialog-Based Online Argumentation: Findings from a Field Ex-
periment. In: Proceedings of the 1st Workshop on Advances in Argumentation in Artificial Intelligence.
pp. 85–99 (November 2017)
[12] Kunz, W., Rittel, H.W.J.: Issues as elements of information systems, vol. 131. Citeseer (1970)
[13] Martınez, D.C., Garcıa, A.J., Simari, G.R.: An abstract argumentation framework with varied-strength
attacks. In: Proceedings of the Eleventh International Conference on Principles of Knowledge Repre-
sentation and Reasoning (KR’08). pp. 135–144 (2008)
[14] Neugebauer, D.: Dabasco: Generating af, adf, and aspic+ instances from real-world discussions. In:
COMMA. pp. 469–470 (2018)