: it 
lffer 
in-

L-

Ie 
.t can 
'remiss 
I up 
:he 
,1em 
: will 
.nc-
in 
of 

:ho.se? 
:a1, " 
lei to 
tI, 
the 
so-

renu -
In, in 
.on 
'as-
sun's 
Ia 
len 
.n 
It 
Ite 
,thout 
that 
oc-

es 

e 
n. 

to 
if 

Ie to 
:ive 
Its 
Ic1u-

:mpt-
tive 
: of 
ound. 
con-

nc-

lining 
ones 
s of 
r by 
mid-
cal 

on. 
ting 
what 
at 
t 
must 

enmine the relation of the premisses to the 
wr1d which occasioned them. To call the 
difference between these two ways to examine 
reasoning the difference between formal and 
informal logic might be useful. But dOing so 
wuld oversimplify and risk another trouble-
some distinction or variant of an old one. 
Md with 2,OOO-odd years behind it our craft 
already has distraction enough • 

* * 

"What Type of AJ:gtDllent 
is an Ad'Verecund~am?" 

* 

John Woods (University of Lethbridge) 
Douglas walton (University of Winnipeg) 

Elsewhere l we have stressed the need, in 
teaching informal logic, to include in ~he 
logical repertoire the skill of discernl.ng 
the ~ of argument that the student is to 

, evaluate. 'For if there is more than one type 
of argument, as we believe, the correctnes~ 
or incorrectness of an argument may vary wl.th 
the factor of type. For example, if there 
are inductively correct arguments, some of 
them (perhaps even all of them) may be deduc-
tively incorrect (invalid). consequently, . 
neglecting this type of distinction could 
spawn many a fallacy. For example, a ~ystem­
atic sophist might take one's correct l.nduc-
tive arguments and rule them deductively in-
correct, efqo bad arguments. For all their 
deductive l.ncorrectness they may be perfectly 
good arguments, taken as what they were meant 

• to be, i. e. inductive arguments. Thus the 
sophist's ploy is based on a true premiss. of 
deductive incorrectness, but it is a sophl.s-
tical refutation because it equivocates on 
the factor of type. 

This factor of type is particularly crit-
ical in teaching the evaluation of arguments 
ad verecundiam. It has sometimes been 

tiliought reasonable that appeals to authorit¥ 
CM be a legitimate type of argument--that l.8, 
not always fallacious--but rather fallacious 
only given that certain conditions of the 
appropriateness of the appeal fail to 
obtain. 2 Even so, one may ask--what type of 
argument is involved? Hamblin (1970, p. 218) 
suggests that we could start from the valid 
argument, "Everything X says is true, X lIaid 
that p, therefore p," and expect to find 
weaker but still not fallacious forms of 
argument where premisses of the form "X is an 
authority on facts of type so-and-SO" lend 
lome support to p. Salmon (1963, p. 64) 
asserts however that the appeal to authority 

5 

is not deductively valid, for the premisses 
could be true and the conclusion false--no 
authority, by these lights, is infallible or 
omniscient. Rather, according to Salmon the 
appeal to autho:r:'ity may be inductively cor-
rect if it has this form: "The vast majority 
of statements made by X concerning subject S 
are true. p is a statement made by X con-
cerning subject S; therefore p is true." 

Who is right? Is the ad verecundiam a 
type of argument that can-Se either deductive 
or inductive, is it perhaps inductive but 
never deductive as Salmon urges, or could it 
be something else altogether, neither deduc-
tive nor inductive in character? These are 
fundamental questions for anyone who would 
want to find ways of teaching students to . 
identify and evaluate the ad verecundiam. 

Two fundamental characteristics of appeals 
to authority ,should be brought forward at 
this point. First, ad verecundiam, like its 
partner in crime ad hominem, is subject-
based. That is, what one authority X asserts 
may-In general be diff~rent from or even con-
tradictory to ~lat is asserted by another 
authority Y. Second, ad verecundiam is 
subject-mAtter-sensitive. That is, an au-
ority's pronouncement that p may be correct 
or not depending on whether or not the 
subject-matter of p is one in which the pu-
tative authority is indeed a legitimate 
expert. Since neither the subject-based or 
subject-mAtter-sensitive characteristics are 
true of the standard or classical approaches 
to the logic of either deductive implication 
or inductive conditionals, it seems reason-
able to think that there may be some deeper 
reasons why the ad verecwldiam can be neither 
deductive nor inductive as a type of argument. 
But how could it be proved? 

We would now like to introduce the thesis 
that arguments ad verecundiam could be Df a 
type that is neither inductive nor de~uc­
tive, an,d suggest that the required type is 
that of the plausible inference of Rescher 
(1976). Plausible reasoning comes to bear on 
cases of informational-overdetermination, 
e.g. inconsistency; where we have too much 
information and have to decide what must be 
given up. Charll.cteristic thar~fore of the 
case of plausible reasoning is the less ~han 
total veracity of our pources, for in an 
inconsistent pair of pronouncements, one 
source must be wrona. In this climate, 
neither deductive nor inductive inference is 
a propos, and in fact Rescher proves that the 
required type of argument can b~ neither de-
'ductive nor inductive. 

Here are the essentials of the proofs 
given in Rescher (1976, p. 2ff.). If the 
inference nx (a generally veracious but im-
perfect source) maintains p, therefore p" 
were deductively valid, then so would the 
following inference be deductively valid for 
some other generally veracious but imperfect 
source Y: ny maintains, p, therefore, p.n 
But if both inferences are indeed deductively 
valid then from nx maintains pn and "Y m~in­
tains , pn it follows that PA , P is true. 
Clearly this consequence' is absurd however, 
for merely because authorities maintain 



conflicting pronouncements it hardly follows 
that p"', P is in fact true. The same con-
sequence follows by elementary laws of prob-
ability from taking "X (a generally veracious 
but imperfect source) maintains p, therefore 
p is highly probable" as a correct inference. 3 
Thus Rescher has shown that for an essen-
tially subject-based (for two sources X and 
Y, or greater than two) appeal to authority, 
the type of inference can be neither deduc-
tive nor inductive. In essence, these dis-
proofs reflect the conception that for mul-
tiple authorities that are imperfect and may 
be expected to have conflicting pronounce-
ments, deductive and inductive models of 
inference are "too perfect". Hamblin's and 
Salmon's conceptions of the ad verecundiam 
are too idealized to adequatery represent the 
practice of appeals to imperfect authorities 
whose pronouncements may clash. But con-
fronted by contradiction we mus~ not give up 
--even though deductive or inductive logics 
give no further guidance--but press on to 
resolve the contradiction by means of plausi-
bility theory. 

Now that we have eliminated the deductive 
and inductive models, and identified plau~i­
ble inference as a preferable model for the 
type of argument exemplified by the ad 
verecundiam, it would seem the way is open 
to an analysis of this fallacy. And so in-
deed it may be, but this is not a project we 
shall attempt here. Suffice it to say for 
the moment that as Rescher conceives it 
plausible inference is not subject-matter-
sensitive, so at very least plausibility 
theory will have to be conjoined to a theory 
of the subject-matter content of propositions 4 
in order to be adequate to the full ad 
verecundiam. These refinements aside how-
ever, we are at least in the position now of 
being able to identify one noteworthily in-
sidious form of the ad verecundiam. 

The fallacy we allude to occurs where an 
appeal to authority is construed so strongly, 
or such a lack of specification of its type 
of argument has transpired, that the argument 
is taken to have (a) deductive, or (b) in-
ductive correctness. Yet if the appeal is 
ment to be taken--as it should be generally--
to a less than perfectly veracious authority, 
then its construal as (a) or (b) is falla-
cious. The specific fallacy here lies not in 
the appeal to authority as such, but in the 
spurious escalation of the appeal towards a 
claim to a source of truth that is more 
perfect or infallible than a plausible argu-
ment has any logical right to be. In short, 
this fallacy is to misidentify the type of 
argument. 

This particular error is of course not the 
only way in which an ap~eal to authority can 
go wrong, and elsewhere we have suggested 
that ad verecundiam is an umbrella concept 
for several specific pitfalls of argument 
from authorities. But this particular spe-
cies of the ad verecundiam is an important 
one, we think, in teaching students how to 
confront and deal with the fallacies, be-
cause it underscores the need to take into 
consideration identification of the type of 

6 

argument as a necessary skill of informal 
logic. The first step in attempting to 
adjudicate any allegation that a fallacy has 
been committed is to ask the question "What 
(exactly) is the argument?" Answering this 
question involves more than simply specifying 
a set of propositions--as in the approach of 
formal logic--it includes, among other tasks, 
specification of the type of argument that 
has been advanced. 

Notes 

lsee our article "Fallaciousness Without 
Invalidity?" Philosophy and Rhetoric, 9, 
1976, 52-54, and HFormal-r,Qgic ana the Logic 
of Argument" to be presented at the 6th 
International Congress of Logic, Methodolo~y 
and Philosophy of Science in Hannover, 
Germany, August, 1979. 

2see our article "Argumentum Ad 
Verecundiam," Philosophy and Rhetoric, 7, 
1974, 135-153. ---

3The proof, parallel to the one above, is 
given by Rescher (1976, p. 3). 

4 
For such a theory, the reader should look 

to Douglas N. Walton 'Philosophical Basis of 
Relatedness Logic,' Philosophical Studies, to 
appear. 

References 

C. L. Hamblin, Fallacies, London, Methuen, 
1970. 

Nicholas Rescher, Plausible Reasoning, 
Assen/Amsterdam, Van Gorcum, ~976. 

Wesley Salmon, I~gi9' Englewood Cliffs, 
N.J., Prentice-Hal~, ~ 63. 

discussi on notes 

A NOTE ~N THE "SURPRISE TEST" PUZZLE 

Harry A. Nielsen (University of Windsor) 

A schoolteacher announces to her 
class that there will be a surprise 
test during the followinq week. She 
specifies that by a "surprise test" 
she means one which no one could 
reasonably predict while walking to 
school. Immediately, one of her 
~righter students claims that she has 
contradicted herself. He offers this 
argument: The surprise test could 
take place on Friday, for if there had 
been no test up until Friday, then 
from that fact and the knowledge that 
taere will be a test any student could 

prec 
he 1 
Fril 
bet1 
SaJIIl 

Tha' 
stu 
tha 
Fri, 
and 
die 
tha 
ext 
be 
The 
be 

out 
fol 
the 

hal 
C1E 
SUJ 

thE 
arc 
thi 

It i! 
that is 
evening 
dity sue 
bright i 
went by 
surprisl 
--Sorry 
you can 
That is 
imagin:i: 
thus ru 
puzzle, 
our for, 
time be 
the end 
before 
the tea 
quence 
her sur 

The 
test ex 
to near 
class p 
does no 
minutes 
student 
er came 
could n 
school. 
doesn't 
too muc 
prise i 
lapse c 
show a 
point, 
a posit 
until t 

wit!' 
if what 
tentia] 
take he