The Philosophical Quarterly Vol. 43, JVo.

173
ISSN 0031-8094

THE LOGICAL GRAMMAR OF THE

TRANSCENDENTALS

BY D. P. HENRY

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

That much of scholastic and contemporary logic-tinged philosophy
constitute two facets of a single conversation, sometimes to the point of
uncanny verbal parallels, should by now be past any doubt. Reminders
on this score are readily available on the side of all branches of logic,
logical metaphysics, and philosophy of language. What is lacking,
however, is a clarification of a sample mediaeval nexus between these
hard-nosed disciplines and more speculative-sounding transcendentals
such as the true and the good. I now propose to provide a sketch
extending to both aspects of that nexus.
In the words of Egbert P. Bos, 'This objective presupposes the
possibility of a real comparison and a real discussion with older
theories'. Moody, Nuchelmans, and Hubien, as well as contributors to
the recent Cambridge History ofLater Medieval Philosophy, are among those
who concur. 1 The objective also presupposes certain trite arithmetical
truths about the numbers of parts of speech which our present-day
discourse has available, as compared with those available to the
scholastic author. If the latter exploits at least twenty parts of speech,
and the language of our present-day elucidation can only field at most
eight, then the elucidatory enterprise is doomed to failure from the
outset. This was already very much a mediaeval issue. Boethius, the
'last of the Romans', had pointed out that technical propositions such
as 'Man is a species' could be construed in such a way that 'species' was
taken to be one of the nomina nominum, 'names of names' (whence
'nominalism' as the title of one stance in this controversy). In this
instance the named name is, of course, 'man'. 2 Ockham is one who
adopts and extends this stance, using the same terminology,' with the
result that he denies that propositions with infinitives as their terms
really are the verb-termed forms which they appear to be. 4 So 'to-be-a-
1
Bos p. 3; Moody p. 10; Nuchelmans preface; Hubien p. 14.
2
Migne cols 176-7.
5
Boehner 1957 p. 36.
* Numquam verbum potest esse extremum propositions (Boehner 1957 p. 188).
C The fdicors of The Phihsophital Qiimkrly, 1993.
432 D. P. HENRY

man is to-be-a-rational-animaV, which follows Boethius' indication that

definitions state the 'to-be . . . ' of the sort of thing in question, 3 then
has to be demoted to the status of a two-name proposition, instead of
the two-verb-termed case which it obviously exemplifies. Clearly, a
two-name theory of the proposition would suffice to analyse such
Ockhamist theories, but would not do for those who took advantage of
unrestrictedly enlarged semantic space, and hence could allow more
complex propositional frameworks. In short, plain arithmetic requires
that no nominalist reductionism (in the sense just indicated) must be

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

presupposed if the non-Ockhamists to be mentioned hereunder are to
enter the conversation. Of course Ockham's contributions to our joint
conversation are nevertheless superb in their own region. 6
A list of some notabilia of contemporary philosophy which display a
certain logical or linguistic tinge would surely comprise linguistic
analysis, with its distinction between real (or logical or deep) and
apparent (or merely grammatical or surface) forms, Frege's sense and
reference, the type/token ambiguity, universal grammar, type-dis-
tinctions in general, usage and meaning, artificial languages, and
varieties of set-theory. Yet strands of all these and many more kindred
fields were being debated almost ad nauseam for hundreds of years in the
culture of Western Europe. That the discussion took place in Latin was
both its felicity and its misfortune. Its felicity was that of the common
language at present so much needed in the European Community and
now only exploited in the news from Radio Helsinki. Its misfortune was
that the technicalities, of a sort which nowadays also sometimes tend to
proliferate around the language of 'symbolic' logic, were expressed in
modes of speaking for which the Renaissance classicist pedants could
find no precedent in the corpus of the dead language of the ancients.
Thus were those modes (and the then common conceptual riches
thereto correlated) condemned to an oblivion later confirmed by
pressure from certain retainers of the Anglo-Norman elite whose wealth
was based upon contempt for the former culture. Whence issued
context-ignoring strictures such as those of Hobbes and Locke, aimed
at 'the frivolous use of uncouth, affected, and unintelligible terms'
which made philosophy 'unfit or uncapable to be brought into well-
bred company and polite conversation'. Although this may look like a
servile trahison des clercs, it may also be seen as an echo of Ockham's
nominalist reproofs of other philosophers' ways of talking. 7

5
Definitio est oratio quae uniuscuiusque rei quidem esse designat (Migne col. 1196C).
6
Henry 1984 sec. 2.6.
7
Henry 1984 pp. 125, 178.
© The editors of The Philosophical Quarterly, 1993.
 THE LOGICAL GRAMMAR OF THE TRANSCENDENTALS 433

A preliminary tabulation of a few present-day rediscoveries of the

lost domains may take its cue from Wittgenstein's summary of
'Russell's merit' in Tractatus 4.0031: this merit was 'to have shown that
the apparent logical form of the proposition need not be its real form'.
Given that 'apparent logical form' is often expressed by Russell as being
the 'grammatical form', as contrasted with the 'logical form', one sees
forthwith that here is a revival of a well-marked tradition which
explored the contrast between grammatical form (as dictated by
traditional grammar, overtly based on usage, usus loquendi) and the

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

technical logical form, a contrast which can be made most fully explicit
in terms of the mediaeval logical or 'speculative' universal grammars.
From the 'Roman' Boethius, throughout Anselm of Aosta and
Canterbury, as well as in the thirteenth-century Boethius of Dacia and
beyond, 8 one has exemplifications of this clash, as when the topic is
that of empty apparent names such as 'nothing', nihil. Anselm's
'Nothing taught me to fly' (Monologion XIX) is the direct analytical
ancestor of Lewis Carroll's 'Nobody passed me on the road', in so far as
both are used as reminders of ordinary grammatical inadequacy. 'We
should not', says Anselm, 'let ourselves be entrapped by the impro-
prieties of words which cover up the truth; rather should we seek after
the propriety of truth which is hidden under diverse manners of
expression', 9 and his writings are replete with further proclamations
and exemplifications of this programme. He even goes so far (as we are
to note below) as to countenance what would ordinarily be seen as
grammatical nonsense in order to separate himself from the clogging
weight of usus loquendi. Here also sense (significatio per se) is contrasted
with reference (appellatio, ancestor of the voluminously-treated sup-
positio). Abelard records the distinction between logical and gram-
matical sense,10 and the later Walter Burleigh not only distinguishes in
word-for-word Russellian style between logical and grammatical
subjects [subiectum quoad logicum, subiectum quoad grammaticum), but also
directs attention to an unusual syllogism which turns out to be
uncannily common to them both in this discussion, namely, 'Men exist;
Socrates is a man; therefore Socrates exists' (omnis homo est; Sortes est
homo; ergo Sortes est).u Russell's analysis of'The round square does not
exist' is inferior to Ockham's elucidation of the corresponding 'The
Chimera is a non-being' (Chimera est non-ens) in that the former lacks
that vital existential exponent which the latter provides and which is

8
Migne col. 408D; Henry 1967 4, 5, 6; Pinborg, Roos and Jensen pp. 65-6.
9
Henry 1984 pp. 149-53; Henry 1967 6.232, 6.6.
10
DalPrap. 271.
" Boehner 1955 pp. 41, 54; Russell p. 164.
© The edilors of The Philosophical Quarterly, 1993.
434 D. P. HENRY

precisely what is required for the avoidance of one of the most familiar
versions of Russell's Paradox. 12 In general, the mediaevals had long
shown themselves capable of handling the non-existent tame tigers in
the Kneale-Moore circus, as well as the Meinong-Russell golden
mountain, without any necessity for entanglement with the
theoretically superfluous gyrations of the 'Theory of Descriptions'. The
notion of a singular verb (which is what Russell's account of'definite
descriptions' tortuously exemplifies and exploits) was already in
evidence in the eleventh-century Garland the Computist's Dialectical

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

This same capacity would allow mediaeval philosophers to come to
terms with Heidegger's 'The Nothing noths' without having to dismiss
it as nonsense in Carnap's style.14 As for the varieties of set-theory, these
may be seen in the then on-going 'universals' controversy. In
particular, the concrete set-theory of collective classes (nowadays
newly fashionable as 'mereology') was dealt with extensively and most
ingeniously across the centuries from Abelard onwards. He himself was
heavily involved in dialectical clashes concerning the specification of
'principal parts' of such integral wholes,13 and evidence is available 16
to suggest that this involvement explains as low-grade intellectual wit
the particular physical manner in which Heloise's guardian chose to
avenge the logician's denial that he had married the girl. For the
provision of an example of the quite common use of the type—token
distinction (on this occasion deployed in connection with the Paradox
of the Liar) Sten Ebbesen's edition of an anonymous commentator on
De Sophisticis Elenchis cannot be bettered."
Such precisely identifiable foundations, mediaeval and con-
temporary, are quite indispensable grounds for any settled
understanding of some scholastic treatments of more speculative topics,
such as those which concern the true and the good. Eschewing, as already
described, any nominalistic two-name restrictions, a suitably sym-
pathetic philosophical grammar may be simply but explicitly
adumbrated by leaving on one side the details of the presupposed
background logical language which is categorial (in the sense of
showing by its form the parts of speech of its components) and making
temporary use of an ad hoc English of a mildly stilted sort, supplemented
12
Boehner 1962 pp. 256, 259; Quine p. 90.
13
de Rijk p. 50.
14
For example, Braakhuis (pp. 80-1) shows how Nicolas of Paris defends nihil nihil est in
such a way that it amounts to a true weak inclusion which is a simple and unobjectionable
counterpart of'The Nothing noths'; cf. Henry 1984 pp. 120-2.
15
de Rijk pp. 549-53; cf. Henry 1991 pp. 92-115.
16
Cousin p. 803; cf. Henry 1991 p. 115.
17
Ebbesen pp. 224-5; cf. Henry 1991 pp. 372-82.
© The editors of Thi Philosophical Quarltrh, 1993.
 THE LOGICAL GRAMMAR OF THE TRANSCENDENTALS 435

by dots and asterisks. This logical-grammatical exercise'8 assumes as

primitive parts of speech the name (in the broad classical sense,
comprising adjectives and common as well as proper names) and the
proposition (in the concrete mediaeval sense of token assertive sentence).
Such primitives serve to specify certain forms which are nowadays
known as 'functors', but which were in mediaeval times called
syncategoremata, and concerning which vast treatises abounded. A
functor or syncategorema is an incomplete expression which upon due
completion forms an expression pertaining to a specifiable part of

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

speech.
Thus, imitating the largely article-free mediaeval Latin, one may
fairly obviously, in the first place, see the f u n c t o r ' . . . i s . . . ' as forming a
proposition from two names, in the nominalistically-recognized way, as
instanced by 'Socrates is literate', 'Plato is captain', 'Augustine is good',
and so on. Furthermore, it is possible to have a non-nominalist but non-
Platonic construal of discourse which deals with essences, forms,
natures, or 'quiddities', of which discourse the already-encountered
'Man is a species', along with its theoretical (e.g., definitional)
accompaniments, may do representative duty, but under a less familiar
but more consistent construal. This construal now views the terms of
such discourse as being the completions of, for instance, another type or
category o f ' . . . is . . . ' . For our ad hoc purposes this may be exemplified
by a starred ' . . . is* . . . ' , the completions of which are no longer names,
but rather more akin to verbs or predicates, themselves instances of
another species of functor. (Such completions will also be starred
below.) In other words, the part of speech of such a two-argument
functor is no longer of the plain name-termed sort, but rather that of
that verb-termed level of locution, of which Boethius reminded us
above when speaking of definitions, and which need have no existential
import whatsoever. Drawing upon natural-language examples which
were familiar to the mediaevals in their Latin versions (currens est movens,
laborare est orare, vivere est esse viventibus, and so forth) it is possible to
exemplify this verb-termed sense by means of participles or of the
infinitives which embrace Boethius' infinitized definitional terms and
their similarly categorized accompaniments and consequences. Such
participles or infinitives do duty for the appropriate deep-structural
'many-link' functors which complete this ' . . . is* . . . '. Thus 'being-a-
man* is*forming-a-species*\ or 'to-be-a-man* is* to-form-a-species*', 'being-
a-man* is* being-an-animal*' reflect the play of the 'many-link' functors,
so called, for example, because of the two-directional aspects of the

18
Remotely and ultimately based on Lejewski 1975; cf. Henry 1991, section 10.
© The fdilors of The Philosophical Quarterly, 1993.
436 D. P. HENRY

terms involved; e.g., 'being a . . . ' not only has an intrinsic

incompleteness (here indicated by dots) but may also (as here) itself
enter into extrinsic relations, as when it forms a term completing the
' . . . is* . . . ' . Abstract nouns, betokening as they do essences or forms,
may also be accounted for in this way, with the categorial language in
the end looking after the ultimate structural details. T h i s ' . . . is* . . . '
may be called 'quidditative', given the level at which it enables the
utterance of general truths about how things are, albeit without
ontological commitment to abstract entities of a Platonic sort. T h i s ' . . .

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

is* . . . ', along with the marginally more familiar first-mentioned
nominally-flanked'... is . . . ' , together allow us to use, as we have seen,
two appropriately correlative types of completion, namely the verb-like
and the name-like. Now a part of speech which received its name from
the fact that it was supposed to 'participate' in the nature of both verb
and name is the participle. This allows the 'currens' and 'movens' of the
foregoing example to be construed in a dual fashion (yielding the senses
of the English 'runner', 'running' and 'mover', 'moving' respectively)
and hence as capable of standing as completions of both types o f . . . is
. . . ' . More importantly, their common termination which occurs in the
Latin, namely ' . . . ens' , is susceptible of a like dual construal, and
happens also to be a key metaphysical term used by the mediaevals,
with 'being' as its usual English translation.
These basic facts a b o u t ' . . . ens', along with later efforts to tidy up the
grammar of being (in the works of subsequent scholastics such as
Cajetan and Sylvester of Ferrara) confirm that ' . . . ens' was indeed
generally seen as a functor in philosophical Latin, with an illuminating
two-level interpretation thereof emerging when their distinction
between ' . . . ens ut nomen' and ' . . . ens ut participium' is taken into
account. Seen as a name ( . . . ens ut nomen) this Latin ' . . . ens' is in the
first place a name-forming functor corresponding to the functorial
termination ' . . . er' in ad hoc English, and to the likewise functorial
termination ' . . . ens' ( o r ' . . . ans') of the nominally-interpreted present
participle in Latin. Examples in the two languages would be 'runn-er',
curr-ens, and 'teach-er', doc-ens. These incidentally show how
' . . . ens' in t h i s ' . . . er' sense is capable of forming names from verbs, and
among such verbs it may be useful to count abstract nouns; as Giles of
Rome in his Theoremata exemplifies the case, 'wisdom ensures that its
subject is a wise-er' (sapientia dat suo subiecto quod sit sapiens)}9 Names
thus formed may in turn constitute completions of functors such as the
name-termed existential'... i s . . . ' : e.g., 'the runn-er is a mov-er' is one

19
Hocedez p. 78; cf. Henry 1984 p. 303.
© The editors of The Philosophical Quarterly, 1993.
 THE LOGICAL GRAMMAR OF THE TRANSCENDENTALS 437

interpretation of the 'curr-ens est mov-ens' example. In contrast, when

seen as a participle (... ens ut participium),'. . . ens' may also (following
the indications of a logical preference noted by Garland the
Computist 20 ) at least minimally and initially be parsed as being verb-
like rather than name-like. It thus corresponds to t h e ' . . . ing*' of the ad
hoc English. (These dual correspondences are hence the very minimum
essentials required for any English-speaking attempt to recover the
senses of much mediaeval Latin talk about being.) This verb-style
functor may be used to form completions appropriate to the

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

quidditative ' . . . is* . . . ' . Like the latter, they may be asterisked in the
present informal context as an ad hoc reminder of this quidditative level
of occurrence, as when 'runn-ing* is* mov-ing*' now shows the
alternative sense of the same 'currens est movens'.
Thus are satisfied the demands of high-mediaeval metaphysicians
such as Giles of Rome (especially in his Theoremata), Thomas Aquinas,
and Patrick of Ireland, for one among the senses of being which does not
demand extra entities in the universe over and above the individuals
named by nominal terms at the level of the nominally-termed ' . . . is
. . . ', for instance, and which can at the same time allow true, reality-
orientated, quiddity-talk to go ahead uninhibitedly yet non-
Platonically. Thus armed, we are also in a position to locate precisely
the status (relatively to a presupposed unproblematical present-day
categorial or canonical language) of the way in which such thinkers
treated both good (bonum) and good* (goodness*, bonitas*), and similarly
with other transcendentals such as verum, the true, and unum, the one,
given that we have already drafted a minimum of the indispensable
outline of their correlated and likewise transcendental ' . . . ens', and
' . . . ens*', with their categorial or typological ambiguity. There
happens to be an easily-available and ample treatment of this topic in
question 21 of Aquinas' Quaestiones Disputatae de Veritate, and it is to this
question that allusions are now to be made.
Hence when, in the corpus of that question's first article, Aquinas has
argued that good, like other transcendentals such as unum (one) and
verum (true), involves something over and above the bare notion of
being ( . . . ens), we are now already equipped to appreciate the
illuminating contrast which he makes between the ways in which bonum
and verum perform this addition, looking at the operation in a purely
typological or categorial way {secundum rationem tantum). (Note that here
verum and bonum are being used paronymously, and hence in English
could, at the nominal level, mean 'a good (thing)', 'a true (thing)', 'the

20
de Rijk pp. 48, 71; cf. Henry 1984 p. 140.
© The editors of The Philosophical Quarterly, 1993.
438 D. P. HENRY

good (being)', 'the true (being)', and so forth.) The contrast which he
makes is based precisely on the distinction (explored above) between (i)
the quidditative level alone {secundum rationem speciei tantum) which in
our ad hoc English notation could involve occurrences of functors such as
' . . . is* . . . ' completed by ' . . . ing*'-terms or infinitives, for instance;
and (ii) the contrasting level of the esse (to be . . . ) that an . . . ens, in the
' . . . er' sense, has in reality {in rerum natura), i.e., at the level normally
associated with name-flanked functors such as the p l a i n ' . . . i s . . . ' , the
completing names of which can perform the office of extra-linguistic

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

reference {appellatio, suppositio) to individual-object reality.
Thus an . . . ens* can in the first place be completive {perfectivum: note
the persistent functorial talk) at the quidditative level alone {secundum
rationem speciei tantum). In this way the understanding is completed by
being* in the quidditative sense {perficiturper rationem ends*) and not by
beings catered for at the concrete esse (or nominally-termed) level
mentioned above. (Need it be interjected that we are of course here
quite unmystically skirting the essence/existence distinction?) It is such
a quidditative-level completion (and this alone) which lies behind the
addition otverum* (the true*, i.e., truth*) concerning any . . . ens* ( ' . . .
ing*') in question. From being mere . . . ens*, the latter becomes verum
. . . ens*, with the gap now further fillable by functors such as 'hab-
. . . ', 'signific- . . . ', lcaus- . . . ', and so forth, in accordance with the
analogical context in question. Evidently 'hav-ing* . . . . ' , 'signify-ing*
. . . ' and 'caus-ing* . . . ' are here corresponding quasi-English verb-
style cases. (At this point we are on the edge not only of Aquinas'
often-mentioned doctrine of analogy, descendant of earlier discussions
of paronymy, but also of yet another large and comparatively un-
known or misunderstood doctrine, characteristically mediaeval, that
of Veritas rei, truth of the thing, or ontological truth. Question 1,
articles 3 and 4 of De Veritate give an inkling of the scope of these.
Most important is the associated tenet that the intellect in question
in Aquinas' oft-quoted definition of truth as the 'adequation of
intellect and thing' (which leads usually to a simpliste classification of
Aquinas as a 'correspondence-theorist of truth') is primarily divine, not
human. 21
In the second place (and it is here that the full force of Aquinas'
illuminating contrast is being exerted) completion by ' . . . ens' (he
continues) need not thus be confined to the quidditative level, but may
extend also to the esse in rerum natura (concrete existence), allusion to

21
As a close inspection of questions 16 and 17 of Part I, and question 109 of Part II—II,
in his Summa Theologica, will soon confirm.
© The edilorj of Thr Philosophical Quarterly, 1993.
 THE LOGICAL GRAMMAR OF THE TRANSCENDENTALS 439

which may involve name-flanked two-argument functors, as already

noted. Here, from being mere . . . ens (in the context of a part of speech
characterizable by the English ' . . . er') the latter becomes bonum . . .
ens, 'good . . . er', now understood nominally and not exclusively (as in
the previous verum case) as a verb, provided that the completion
relates to concrete fulfilment or preservation (i.e., is «reaf-orientated
with respect to that concrete object which it completes or accomplishes,
in quantum unum ens est secundum esse suum perfectivum alterius et conservativum
habet rationemfinisrespectu illius quod ab eo perficitur). So, by analogy with
the healthy as either the hav-er, or the complet-er, or the keep-er or

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

the signifi-er of health, we now have the good (object) as the good
either qua end-er (the . . . ens which is the inspiring purpose) or qua
means-er (the . . . ens which is the useful means) or qua the end-
consequence-er (the . .. ens which is good as a consequence of the good
end). 22
That verum* stands contrastingly at the quidditative level alone is
confirmed by the reply to the third objection of article 3 in the question
21 which is now being scrutinized. Thus 'the nature of the true* issues
forth from the species* in so far as it is understood as it really is*' (ratio
... veri* ex ipsa specie* consurgit,prout est intellecta sicuti est*). On the other
hand bonum, 'good', is yet again shown also to extend (in contrast
to this categorial restriction of verum*, 'true*') to the nominally-
termed level when the question of the theoretical ranking of the
transcendentals among themselves is discussed in the corpus of this
article 3:

If one looks at the true and the good considered in themselves, then the
true* is by nature prior to the good, since it is completive of something
at the species*-level [i.e. as embraced by theory]; the good, how-
ever, performs its completive function not only at the species*-level,
but also in respect of the existence that it has in things. On this account
the nature of good covers a wider [categorial] range than does the
nature of true*, and has the status of something extra, over and above
the other. In this sense the good presupposes the true*, whereas the
true*, in its turn, presupposes the one*. This is because the nature of
the true* is fulfilled by intellectual grasp; every item* is intelligible,
however, to the extent that it is one*; he who does not understand the one*
does not understand any thing*. Hence the order of these transcendental
names, if they are looked at in respect of themselves, is as follows: after
22
Sic ergo primo et principaliter dicitur bonum ens perfectivum alterius per modumfinis;sed
secundario dicitur aliquid bonum quod est ductivum infinem; prout utile dicitur bonum; vel natum est
consequi finem; sicut et sanum dicitur non solum habens sanitatem, sed perficiens et conservans et
significans (De Veritate q.2\, art. \,c).
© The editors of The Philosophical Quarterly, 1993.
440 D. P. HENRY

being comes one*, then the true*, and finally, after the true*, comes the
good.23

The mention of the transcendental one* here is easily accounted for.

The quidditative-level ' . . . is* . . . ' may be used as the basis for the
definition of a non-concretely-existential correspondingly quidditative
'there exists* exactly one . . . ' which (in contrast to what holds at the
nominally-termed level) turns out to be a thesis for all appropriate
completions. Hence indeed, unum*, 'one*', does initially follow from . . .

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

ens*, ' . . . ing*', 'being*', exactly as Aquinas implies. Giles of Rome's
superb Theoremata embody a much fuller version of such inter-
connections. 24
Growing familiarity with the dual senses of' . . . ens' ( ' . . . ing*' and
' . . . er'), and the light which that familiarity is already throwing on the
logical-grammatical status of good and of the other transcendentals,
facilitate our dealing with the prima facie robust-sounding fourth
supporting argument for the thesis that all things are good by primal
goodness (omnia sunt bona bonitateprima: question 21, article 4, argument
4). This argument claims that if one allows every creature to be
described as good (in a secundum quid, derivative sort of sense, as covered
in article 5), then one can ask whether such a creature is good by some
self-inherent good or only by a primal good. If the first applies, then
that self-inherent good will itself be good; but this in turn is either self-
based (i.e., it is good by its own good) or based on another. If it is self-
based, then it is itself primal good. If it is not self-based, then in order to
avoid an infinite regress, one must eventually arrive at some self-based
primal good whereby the originally mentioned good is good. In either
event, therefore, creaturely good is good-by-a-first-good.
As the corpus of the cited article shows, Aquinas is well aware that the
Platonic 'Good' lurks behind such arguments, and that this can indeed
be Christianized by identification with God. Nevertheless, in response
to this fourth argument he resists any such identification, and does so in
terms which may immediately be relativized to our logical-gram-
matical background. Indeed, he plunges the reader forthwith into a

23
Considerando ergo verum et bonum secundum se, sic verum estprius bono secundum rationem, cum
sit perfectivum alicuius secundum rationem speciei; bonum autem non solum secundum rationem speciei,
sed secundum esse quod habet in re. Et ita plura includit in se ratio boni quam ratio veri, et se habet
quodammodoper additionem ad ilia; et sic bonum praesupponit verum, verum autem praesupponit unum,
cum veri ratio ex apprehensione intellectus perficitur; unumquodque autem intelligibile est in quantum est
unum; qui enim non intelligil unum nihil intelligii Unde istorum nominum transcendentium talis est
ordo, si secundum se considerentur, quod post ens est unum, deinde verum, deinde post verum bonum (De
Veritate q.2\, art. 4, c).
24
Henry 1984 pp. 180, 295-316.
© The editors of The Philosophical Quarterly, 1993.
 THE LOGICAL GRAMMAR OF THE TRANSCENDENTALS 441

stream of discussion which stretches back historically, and which has as

its central linguistic identifying symptom the admission (or rejection, as
the case may be) of the interpredication of abstract and concrete forms
of the name (in the broad sense of'name', noted above). Although he
does not mention the fact, it was the aforementioned Anselm who had,
in his De Grammatico, daringly extended this type of interpredication
outside the already existing and quite general theological
countenancing thereof (as in Deus est Deltas, 'God is Godness'). He thus
has cases likegrammaticus* est* grammatica*, 'literate* is* literacy*', and

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

album* est* albedo*, 'white* is* whiteness*'. As the starred functors and
terms here indicate, our logical grammar accounts for these apparent
grammatical oddities (even more odd in the Latin than in the English)
by displaying how Anselm used them as signals that an ascent is taking
place to the verb-termed quidditative level; this is done in order to
elucidate the manner in which such adjectival names (in this case) had
each a verb-participial functorial sense which left the way open for their
specifically varied occasions of reference (particular horses or swans, for
instance, in the case of'white'). 25
We are in fact by now well equipped to appreciate the point of
Aquinas' response. It turns out that, like John of Salisbury (Metalogicon
HI/2) and Dionysius the Areopagite, the second of whom he adduces,
Thomas is against any extensive interpredication of the sort thus
initiated by Anselm. Where the latter used the possibility of such
interpredication to betoken an ascent to the quidditative, Aquinas uses
it to sort out theses from non-theses at an identical quidditative level.
First Thomas justifies his restrictions by distinguishing between
transcendental 'forms' and specific 'forms' (where 'form' is used in the
quite unexceptional sense of'quiddity'). Where transcendental forms
are concerned, and (we may add) as John of Salisbury admitted,
interpredication of abstract and concrete is in order, i.e., 'goodness* is*
good*', 'oneness* is* one*', 'beingness* is* being*', and so forth.
Where, however, more particular (i.e., 'specific') quiddities are in
question, such interpredication is not admissible. Hence all the
scandalously ungrammatical Anselmian extra-theological admissions,
such as grammaticus* est* grammatica*, are now expressly vetoed.
The situation is delineated in more detail by Thomas as he goes on to
explain this relatively surface-structured distinction by reference to its
underlying theoretical bases. He first states his well-known thesis that
the first intellectual apprehension is of being* (in the ' . . . ing*' sense, of
course). He then rightly uses this easily accepted postulate for the
25
Henry 1984 pp. 160-73.
© The editors of Tki Philosophical Quarterly, 1993.
442 D. P. HENRY

evolution of further theorems. Thus it is the case that, no matter what,

essentia* ( . . . ing-ness*) is theoretically understood as . . . ens* ( . . .
ing*); in consequence beingness* is* (be)ing*, whiteness* is*
(be)ing*, and (relevantly to our subject) goodness* is* (be)ing*. His
' . . . et sic de aliis' ( ' . . . and so on in other cases') is his counterpart of the
universal quantifier which one would expect to be appropriate to such
a general thesis.
But (he continues) there are other transcendental notions which (as
we have seen) are bound up with . . . ens ( ' . . . ing*' o r ' . . . e r ' ) , i.e., one,

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

good, and (to some extent) true. Hence these, like being, are sufficiently
generic to be likewise predicable at the quidditative level of no matter
what appropriately essential notion. Hence not only do we have the
interpredication of oneness* and one*, goodness* and good*, as well as
trueness* and true*, but we may also conclude that whiteness* is*
one*, is* good*, is* true*, for reasons broached briefly in the foregoing,
and so on for all abstract or quidditative-level subject-terms, be they
generic or specific. In the light of the roughly explicit logical grammar
of our analysis, this continuation of Aquinas' response is still quite
intelligible. It is indeed the case that each of 'goodness* is* one*',
'trueness* is* one*', or even 'oneness* is* one*' follows in accordance
with the quidditative-level thesis expressed above in terms of an ad hoc
approximate English as 'there exists* exactly one . . . ', with the space
being tillable by abstract nouns or their appropriately quidditative
counterparts. It is precisely because of the generality of what is being
said here that 'whiteness* is* one*' likewise follows. Here once again,
therefore, we know exactly where we are in the logical grammar of our
expressions, which turn out (at least in the case o f . . . is* one*') to be
nowadays derivable as theses or instances thereof, and to be simply
expressive of the uniqueness of any given duly specified abstract set of
things. Having intuited such theses, Aquinas is now characteristically
extending them in respect of all his transcendentals (at the predicate-
term end of the theorem), leaving the way open for total generality,
embracing, among other items, the transcendentals themselves, at the
subject-term end. We hence obtain not only the comparatively prosaic
theses overtly mentioned thus far, but also the more poetic-sounding,
but still sense-making and intelligible, 'truth* is* good*', 'goodness*
is* true*', and so forth.
However, Aquinas continues, the interpredications of abstract and
concrete pseudo-names which are involved in the generalizations just
broached fail when a non-transcendental abstract term is the subject,
and its concrete counterpart the predicate. Hence 'whiteness* is*
white' and the like are excluded. He explains this exclusion in a way
© The editors of The Philosophical Quarterly, 1993.
 THE LOGICAL GRAMMAR OF THE TRANSCENDENTALS 443

consonant with our own non-starred version of'white' here, in that he

makes it plain that (unlike Anselm) he would see this 'white' as a
nominal expression and hence by his, our, and even Anselm's
presupposed logical grammar, not a fit completion for a quidditative-
level ' . . . is* . . . '. As we would broadly expect, the presupposed
canonical language requires verb-like, rather than nominal,
completions in such a situation, hence the exclusion of the nominally-
interpreted 'white'.
In contrast, as Aquinas rightly goes on to say, from the point of view

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

of natural l a n g u a g e , . . . ens ( ' . . . ing*' o r ' . . . er'), need not be taken to
be subject to any such restrictions, since it can adapt itself to all
categorial levels. Thus in the first place something is . . . ens (in the ' . . .
er' sense) because it subsists of itself. Here we are concerned with
nominal-level expressions formed from substance-names, e.g., from
'The Socrates-man is the Socrates-man' we can derive 'The Socrates-
man is an . . . ens ( . . . er)'. This mixture of natural languages for local
purposes indicates clearly how this name-forming ' . . . er' is now in
question, and the categorial expression would be plain enough on this
score: we are here dealing with a name-completed ' . . . is. . . ' .
Next, he continues, something is an ' . . . ens' because it is a
theoretically available basis for the self-subsistents of the first case,
being commonly called a 'form' ( . . . estprincipium subsistendi, utforma).
We are now witnessing the ascent to the quidditative-level ' . . . ens*'
( i . e . , ' . . . ing*') of which sufficient has already been said. Then follow
other examples of categorial varieties of. . . ens, i.e., the disposition of a
subsistent, such as a quality, or the privation of a disposition of a
subsistent, such as blindness (dispositio subsistentis, ut qualitas,. . . privatio
dispositionis subsistentis, ut caecitas). These cases may be analysed
quidditatively (in the first case) or either quidditatively or at the
nominal level (this last being in Ockham's fashion) in the second case.
The point of all these glimpses of quite irrefutable and indispensable
metaphysics is to lead towards a decision concerning the contention
that creatures are good by a primal good (in a Platonic or quasi-
Platonic fashion). From his theses Aquinas first draws further
reminders which converge on the quidditative level. Indeed, they take
their start from the essentia* est* . . . ens*, 'being-ness* is* (be)ing*',
which has already been broached above:

And hence when we assert 'being-ness* is* (be)ing*', if we went on to

derive, 'hence it* is* by some (be)ing*, be it itself or some other', this
derivation is not sound, since the original assertion did not involve'... is a
being' in the same way as something subsisting of itself is a being [i.e., at
© The cdilors of Tht Phihmphiml Quarterly, 1993.
444 D. P. HENRY

the nominal'... er' level] but rather as [already] that by which something
is [i.e., already at the quidditative ' . . . ing*' level]. Hence it is not
appropriate to ask [yet again] how that beingness* itself is* by something,
but rather in what manner something else is [i.e., at the nominally-named
' . . . er' level] by that beingness.
Likewise, when it is asserted 'goodness* is* good*', then 'good*' is not
asserted to be an attribute of goodness* considered as an attribute of a
subsistent being [i.e., we are not here at the nominal ' . . . er' level] but
rather 'good*' is here asserted so as to convey that it is that by which
something is good [i.e., that it is at the quidditative ' . . . ing*' level].

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

Hence it is not appropriate to enquire [as is done in the argument under
discussion] whether goodness* is* good* by its own goodness or by some
other, but rather we should enquire whether it is by that goodness* itself
that a thing is rendered a good [' . . . er'] distinct [categorially] from its
own goodness* ['good-ing*'], as is the case with creatures, or whether it is
a good* thing* which is identical with its own goodness*, as is the case
with God.26

Confined as he is to the limitations of a particular natural language,

Aquinas is here excellently expressing, to the limits of that language's
capacity, the categorial distinctions which are obviously required in
order to make his point about the grammar oigood. Thus in the first of
the three paragraphs just cited, he makes clear the status of being* at the
quidditative level. The essences ('beingnesses') which are the topic of
discourse at this level are shown by him as being themselves not entia,
beings at the nominal' . . . er' level, but rather entia-quo*, beings*-by-
which . . . , thus employing one more very ingenious way of calling
attention to the functorial, verb-like ' . . . ing*', nature of the terms of
the discourse, as opposed to name-based references to beings in an ' . . .
er' sense comprising the Johnsonian kickables. Of such concrete objects
there is a sense in which one may ask 'by what* is it what-it-is?', the verb-
like functorial reply to which takes us to the quidditative. And plainly,
having thus achieved the quidditative, one cannot then in the same sense
ask of one of the entities* covered by its discourse, 'by what* is* it*
what-it*-is?'. Present-day categorial language can without any
difficulty confirm all these Aquinate intuitions. 27 Then, given Aquinas'
26
Et ideo cum dicimus: essentia est ens, siprocedatur sic: ergo est aliquo ens, vel se vel alio, processus
non sequitur, quodnon dicebatur hoc modo esse ens, sicut aliquid subsistens in esse suo est ens, sedsicut quo
aliquid est. Unde non oportet quaerere quomodo ipsa essentia aliquo sit, sed quomodo aliquid alterum sit
per essentiam. Similiter, cum dicitur bonitas bona, non hoc modo dicitur bona quasi in bonitate subsistens,
sed hoc modo quo bonum dicimus Mud quo aliquid bonum est. Et sic non oportet inquirere utrum bonitas
sit bona se bonitate vel alia, sed utrum ipsa bonitate sit aliquid bonum quod sit alterum ab ipsa bonitate,
sicut est in creaturis; vel quod sit idem cum ipsa bonitate, sicut est in Deo {De Veritate q.2\, art. 4, ad
4).
27
Henry 1984 sees 3, 4.
© The editors of The Philosophical Quarterly, 1993.
 THE LOGICAL GRAMMAR OF THE TRANSCENDENTALS 445

characteristic and already-noted alignment of... ens with bonum... and

the other transcendentals, there is no problem in confirming the
intelligibility of the logical grammar of this whole pattern of restrictions
on ascents or regresses such as those which the argument in question has
proposed or threatened.
The relevance to the question ofgood is hence by now plainly obvious.
The veto on the extension of the 'by what* is i t . . . ?' question, beyond
its original range of application at the level of concrete nominally-
named entities, applies as much to goodness as to other quidditative-
level notions, as the third paragraph of the last quotation shows. That

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

same paragraph then changes the slant of the enquiry to the contrast
between the creaturely case (wherein the creature is not identical with
its own good*) and the divine case (wherein there is identity, not only
with goodness*, but with all other attributes*, and with esse (or esse*) as
well). This is, of course, a further salient instance of the analogical (i.e.,
type-discrepant or category-discrepant) senses of 'good' to which
attention had been called in the corpus of article 4 of that question 21 to
which the argument we have been examining is annexed.
The particular instances cited above, and myriads like them, convey
the ease with which the highly linguistically-aware mediaevals handled
the functorial construal described. That ease reflects the manner in
which their heavily inflected Latin's notion of incomplete 'stems' or
'roots' readily gives rise to the functorial concept.28 The largely
functorial character of the present-day categorial language pre-
supposed in the elucidation provided above hence ensures that it
constitutes a most sympathetic medium for the interpretation of areas
such as the nexus described. Indeed, the constantly avowed but equally
constantly unachieved mediaeval ideal of precise derivation of meta-
physics from first principles may thus at last be realized in the manner
made possible by that contemporary philosophical logic which has so
much in common with scholastic thought.

University of Manchester

REFERENCES
Boehner, P. (ed.) 1955: Walter Burleigh: De Puritate Artis Logicae, Traclatus Longior, etc. (St
Bonaventure, N.Y.: The Franciscan Institute).
(ed.) 1957: William Ockham: Summa Logicae, Pars Prima (St Bonaventure, N.Y.: The
Franciscan Institute).

28
Henry 1991 pp. 11-12. Nicolas of Paris (in Braakhuis) is a supreme master of
functorial awareness and application.
© The edicors of The Philosophical Quarterly, 1993.
446 D. P. HENRY

(ed.) 1962: William Ockham: Summa Logicae, Pars Secunda el Tertia Prima (St Bonaventure,
N.Y.: The Franciscan Institute).
Bos, E.P. (ed.) 1983: Marsilius of Inghen: Treatiseson the Properties of Terms (Dordrecht: Reidel).
Braakhuis, H. A.G. 1979: De 13de Eeuwse Tractaten over Syncategorematische Termen, Deel II: Uitgave
van Micolaas van Parijs' Sincategoreumata (Meppel: Krips Repro).
Cousin, V. (ed.) 1859: Petri Abaelardi Opera, Tomus Posterior (Paris: Aug. Durand).
Dal Pra, M. (ed.) 1969: Pietro Abelardo: Scritti di Logica (Firenze: La Nuova Editrice).
de Rijk, L.M. (ed.) 1956: Petrus Abaelardus: Dialectica (Assen: Van Gorcum).
(ed.) 1959: Garlandus Compotista: Dialectica (Assen: Van Gorcum).
Ebbesen, S. (ed.) 1977: Incerlorum Auctorum, Quaestiones super Sophisticos Elenchos (Copenhagen:
G.E.C. Gad).
Henry, D.P. 1967: The Logic of Saint Anselm (Aldershot: Gregg Revivals).

Downloaded from http://pq.oxfordjournals.org/ at University of Manitoba on June 14, 2015

1984: That Most Subtle Question (Quaestio Subtilissima): the Metaphysical Bearing of Medieval
and Contemporary Linguistic Disciplines (Aldershot: Gregg Revivals).
1991: Medieval Mereology (Amsterdam: Gruner/Benjamins).
Hocedez, E. (ed.) 1930: Giles of Rome: Theoremata de Esse el Essentia (Louvain: Museum
Lessianum).
Hubien, H. (ed.) 1976: loannis Buridani Tractatus de Consequentiis (Louvain: Publications
Universitaires).
Lejewski, C. 1975: 'Syntax and Semantics of Ordinary Language', Papers of the Aristotelian
Society, Suppl. Vol. 49.
Migne, P. (ed.) 1891: A.M.T.S. Boelhius: Opera Omnia, Tomus Posterior (Paris: Series: Patrologia
Latina).
Moody, E.A. 1953: Truth and Consequence in Medieval Logic (Westport: Greenwood).
Nuchelmans, G. 1973: Theories of the Proposition: Ancient and Medieval Conceptions of the Bearers of
Truth and Falsity (Amsterdam: North Holland).
Pinborg, J., Roos, H. & Jensen, S.K. (eds) 1969: Boethii Daci Opera, vol. IV: Modi Significandi
(Copenhagen: G.E.C. Gad).
Quine, W.V. 1953: From a Logical Point of View (Cambridge, Mass.: Harvard UP).
Russell, B. 1919: Introduction to Mathematical Philosophy (London: Allen & Unwin).

© The edilors of Thr Philosophical Quarterly, 1993.