Barnes Jonathan. Truth, etc

Подождите немного. Документ загружается.

110 Predicates and Subjects

Bugles sang

you will get

Sang bugles,

or something similar; and such items are ill-formed. (To be sure, you might

meet them in a piece of English; but then you would read them as poetic

inversions.) Again, if you replace the verb by a name or the name by a verb

you get ungrammatical nonsense, such as

Bugles trumpets

Blared sang.

But if you interchange the subject and the predicate of a proposition,

the result—whether true or false—is well-formed. Likewise, if you replace

the subject by a predicate—by an item which is a predicate in some

other proposition—or the predicate by a subject—by an item which

is a subject in some other proposition—then the result is again well-

formed.

Sincethatisso,thepredicatein

Bugles sang

is not ‘sing’ or ‘to sing’ or ‘sang’. But then what is it? In the case of a sentence

Hopes are dupes,

the question is readily answered: the two terms are ‘hope’ and ‘dupe’. They

are linked by the copulative ‘is’ which itself is not a term. The terms are

homogeneous inasmuch as

Dupes are hopes

is well-formed. So the question may be put like this: How are we to ﬁnd in

Bugles sang?

something like the copulative ‘is’?

The traditional answer to that question is this: The copulative ‘is’ is

already present in the sentence—potentially. Thus Ammonius remarks, of

an Aristotelian thesis about negation, that

he shows this in the case of propositions which have ‘is’ potentially—for example, in

the case of ‘A man walks’, he analyses ‘walks’ into the participle and ‘is’, and he leaves

us to deduce that if in the case of the analysed proposition which has ‘is’ actually

we make the negation by tying the denial to ‘is’, then in the case of the proposition

Names and Verbs in Aristotelian Logic 111

which has ‘is’ potentially we shall have to tie the negative particle to the part of the

proposition which contains ‘is’.

(in Int 222.18–24)³⁵

Aristotle actually said that

there is no difference between saying that a man walks and that a man is walking.

(Int 21b9–10)³⁶

Aristotle’s phrase ‘there is no difference’ is vague; but the later tradition not

unreasonably took him to mean that ‘A man is walking’ makes actual or

explicit what is merely potential or implicit in ‘A man walks.’

There is no reason to question the equivalence which Aristotle and his

followers proclaim—although it is perhaps worth insisting that the claim

does not concern the English continuous present (for which, of course, it

would be quite false) but rather a certain Greek paraphrastic idiom. So (the

Greek for) ‘sang’ means the same as (the Greek for) ‘was singing’; and in

general ‘verbs’ means the same as ‘is verbing’.

Aristotle makes the point again in the Metaphysics:

There is no difference between ‘A man is ailing’ and ‘A man ails’, nor between ‘A

man is walking’ (or ‘cutting’) and ‘A man walks’ (or ‘cuts’); and similarly in the other

cases.

(Met 1017a27–30)³⁷

Alexander comments thus:

He states that ‘A man is ailing’ signiﬁes nothing other than ‘A man ails’—i.e. the ‘is’,

which is constructed with illness signiﬁes nothing other than the holding of illness.

Similarly in the case of ‘walking’, it signiﬁes the holding of walking, in the case of

‘cutting’, it signiﬁes the holding of cutting, and likewise in all cases. For as he said in

³⁵ τοῦτο οὖν καὶ ἐπὶ τῶν δυνάμει τὸ ἔστιν ἐχουσῶν οὕτως ἔχον ἐπιδείξας, οἷον τῆς

ἄνθρωπος βαδίζει, τῷ ἀναλῦσαι τὸ βαδίζει εἰς τὴν μετοχὴν καὶ τὸ ἔστι καὶ καταλιπεῖν

συλλογίζεσθαι ἡμῖν ὡς εἰ ἐπὶ τῆς ἀναλελυμένης προτάσεως καὶ ἐνεργείᾳ τὸ ἔστιν ἐχούσης

αὐτῷ τῷ ἔστι συμπλέκοντες τὴν ἄρνησιν ποιοῦμεν τὴν ἀπόφασιν, δεήσει καὶ ἐπὶ τῆς κατὰ

δύναμιν αὐτὸ ἐχούσης τῷ περιέχοντι αὐτὸ μορίῳ τῆς προτάσεως τὸ ἀποφατικὸν μόριον

συμπλέκειν.

³⁶ οὐδὲν γὰρ διαφέρει εἰπεῖν ἄνθρωπον βαδίζειν ἢ ἄνθρωπον βαδίζοντα εἶναι.

³⁷ οὐθὲν γὰρ διαφέρει τὸ ἄνθρωπος ὑγιαίνων ἐστὶν ἢ τὸ ἄνθρωπος ὑγιαίνει, οὐδὲ τὸ

ἄνθρωπος βαδίζων ἐστὶν ἢ τέμνων τοῦ ἄνθρωπος βαδίζει ἢ τέμνει· ὁμοίως δὲ καὶ ἐπὶ τῶν

ἄλλων.

112 Predicates and Subjects

the de Interpretatione, in itself it is nothing, but it co-signiﬁes a sort of composition

which cannot exist without the items composed.

(in Met 371.30–36)³⁸

If I say

Bugles sing,

that signiﬁes nothing other than

Bugles are singing,

and there the verb ‘are’ signals that the activity of singing, which the participle

‘singing’ expresses, holds or is predicated of bugles.

But what good does that do us? After all, the analysis in terms of copula

and participle does not introduce the desired homogeneity:

Singing are bugles

—just like ‘Sang bugles’—is either a poetical inversion or nonsense. Well,

in Greek things are different. Or rather, with a little juggling they come to

be different. And after all, a little juggling works wonders in English too.

Translate Aristotle’s analysed proposition not as ‘A man is walking’ but as

‘A man is an item which walks.’ In general, suppose that ‘verbs’ means ‘is

an item which verbs’. Then ‘sang’ means the same as ‘is an item which was

singing’: it is, as it were, an amalgamation of a copulative ‘is’ and the nominal

phrase ‘an item which was singing’. The homogeneity is now guaranteed: any

nominal phrase of the form ‘an item which …’ may function in the subject

place of a sentence and also in the predicate place. After all, bugles are bugles,

and items which sang are items which sang.

To be sure, the copulative ‘is’ requires scrutiny. Is it—for example—a

genuine verb which indicates a time? If not, then what is it? If so, then why

not parse ‘Bugles sang’ as ‘Bugles were singing items’? Why, in other words,

should the copulative connector not indicate both a connection and a time

of connection? That question will turn up brieﬂy in a later context. Here it

can be left to one side.

In the light of that, what is the relation between names and subjects, verbs

and predicates? Or between Aristotelian grammar and Aristotelian logic? A

partial answer to the question might run like this: In the simplest of sentences,

which consist of one name and one verb, the name will be, or will correspond

³⁸ παρέθετο τὸ μηδὲν σημαίνειν ἄλλο τὸ ἄνθρωπος ὑγιαίνων ἐστὶν ἢ τὸ ἄνθρωπος ὑγιαίνει,

τουτέστι τὸ ἔστιν, ὃ ἐπὶ τῇ ὑγείᾳ συντέτακται, μηδὲν ἄλλο ἢ τὴν τῆς ὑγείας ὕπαρξιν σημαίνει·

ὁμοίως καὶ ἐπὶ τοῦ βαδίζων τὴν τῆς βαδίσεως, καὶ ἐπὶ τοῦ τέμνων τὴν τῆς τομῆς, καὶ ἐπὶ

πάντων ὁμοίως. ὡς γὰρ εἶπεν ἐν τῷ Περὶ ἑρμηνείας, αὐτὸ μὲν οὐδέν ἐστι, προσσημαίνει δὲ

σύνθεσίν τινα, ἣν ἄνευ τῶν συγκειμένων οὐχ οἷόν τε εἶναι.

Porphyrean Predicates 113

to, the subject. No more adventurous or less partial answer will be both true

and informative.

Antiquity did indeed offer a wholly general answer—but it is wholly

boring. Ammonius explains that Aristotle uses the word ‘verb’ in three ways:

it may mean pretty well what we and the ancient grammarians normally

mean by ‘verb’; it may designate a present indicative of what we and the

ancient grammarians normally mean by ‘verb’; and it may signify

any expression which makes a predicate in a proposition; so that in this sense

‘beautiful’ and ‘just’ and ‘white’ and ‘animal’, when they are taken as predicated, are

called verbs, which is not so in either of the two earlier senses.

(in Int 53.5–8)³⁹

So verbs and predicates are one and the same thing—provided that you take

‘verb’ to mean ‘predicate’ (and hence classify names as verbs). Whether or

not Ammonius is right to discover such a use of the word ‘verb’ in the text of

Aristotle, it is evident that the discovery can be of no substantive interest.

PORPHYREAN PREDICATES

Aristotelian subjects and predicates are mutually correlative items (in Greek,

they are ἀντιστρέφοντα), names and verbs are not. A subject is a subject for

a predicate and a predicate is a predicate of a subject. A name is not a name

for a verb and a verb is not a verb of a name. Again, a predicate is a predicate

of a subject in a proposition: a verb is not a verb of a name in a sentence—it

is a verb full stop. (Names, of course, are relational items inasmuch as a name

is a name of something. So they have their correlatives; but the correlative of

a name is the item which it names, not a verb with which a sentence may

associate it.) That correlativity not only distinguishes subjects and predicates

from names and verbs: it also distinguishes Aristotelian predicates from Stoic

predicates (and from the predicates of contemporary logic).

So subjects and predicates are relational items, and predication is a two-

placed relation: ‘x is predicated of y’. Subjection, in that case, is also and of

course a two-placed relation: ‘x is subjected to y’. Subjection is the converse

of predication: x is subjected to y if and only if y is predicated of x. A

³⁹ ... ἢ πᾶσαν φωνὴν κατηγορούμενον ἐν προτάσει ποιοῦσαν, ὥστε κατὰ τοῦτο τὸ

σημαινόμενον τὸ καλὸς καὶ δίκαιος καὶ λευκὸς καὶ ζῷον, ὅταν κατηγορούμενα ληφθῇ, ῥήματα

λέγεσθαι, ὅπερ κατ᾿ οὐδέτερον ἦν τῶν προτέρων σημαινομένων.

114 Predicates and Subjects

predicate is whatever is predicated of something—that is to say, ‘x is a

predicate’ means ‘x is predicated of something’; and a subject is whatever is

subjected to something—‘x is a subject’ means ‘x is subjected to something’.

(In the same way, ‘x is a parent’ and ‘x is a child’ mean ‘x has some offspring’

and ‘x is offsprung from something’.) It is sometimes said that you cannot

coherently speak of subjects and predicates simpliciter —that you cannot ask,

say, if the term ‘llama’ is a predicate or not. After all, a subject is always a

subject of something, and a predicate a predicate of something. That is true;

but it is a half truth: you can coherently ask of an item if it is a predicate,

or a subject, just as you can coherently ask of an item if it is a parent, or

achild.

‘x is predicated of y’ is a passive construction. In speaking of predication

Aristotle and his successors most often use the verb ‘κατηγορεῖν’, in the pass-

ive. It is followed by a genitive, or by ‘κατά’ + genitive: ‘τὸ Α κατηγορεῖται

(κατὰ) τοῦ Β’. There are other locutions; and often enough, especially in

technical expositions, the verb is elided, so that ‘τὸ Α τοῦ Β’means‘Ais

predicated of B’. (In speaking of subjection the Peripatetics use ‘ὑποκεῖσθαι’,

which takes a dative: ‘τὸ Α ὑποκεῖται τῷ Β’. Other formulas are found; but

they are rare—and in fact ‘ὑποκεῖσθαι’ is itself far rarer than ‘κατηγορεῖ-

σθαι’.) The verb ‘predicate’ also has an active use, in Greek as in English.

Thus you may say that a sentence, or a proposition, predicates one item of

another; and you may say that a speaker, or a thinker, predicates something

of something. Very roughly speaking, a sentence or a speaker predicates x

of y if and only if what it or he says is true if and only if x is predicated

of y. And we may suppose that a speaker or thinker predicates x of y if

andonlyifheproducesorentertainsasentenceorapropositionwhich

predicates x of y. A predicable, you might then say, is something which

a sentence or a speaker might predicate of something, and a subjectible

is something which a sentence or a speaker might subject to something.

In Peripatetic logic, an item is a predicable if and only if it is a subject-

ible. That is another way of expressing the homogeneity of subjects and

predicates.

If predication is a two-place relation it invites a few elementary questions:

Is it a reﬂexive relation? Is it symmetrical? Transitive? And so on. If you

can’t answer such questions, then you haven’t got the second idea about the

relation. But in the case of predication the matter is better postponed.

For ﬁrst, I want to raise the following question: What sorts of things are

subjects and predicates? What are the relata of the relation of predication? If

Porphyrean Predicates 115

x is predicated of y, then what kind of items must x and y be? The preceding

pages have vacillated, and the matter demands examination.

It was much discussed in late antiquity, and it was discussed in a particular

context. For the term ‘predicate’ ﬁrst met the student philosopher in the

pages—or more precisely, on the title-page—of the ﬁrst component of the

Peripatetic Organon. The title under which that little essay generally goes in

English, namely ‘Categories’, is not a translation but a quasi-transliteration of

the standard Greek title; and that title, ‘Κατηγορίαι’means‘Predications’.

(Latin writers generally used ‘Praedicamenta’.) As ancient scholars saw things,

the words ‘subject’ and ‘predicate’ were technical terms of Peripatetic logic. To

be sure, the words in question were not Aristotelian neologisms; for Aristotle

was not a great neologizer. But Aristotle had given them new senses, and the

senses needed explanation. No doubt the terminology had ventured beyond

the borders of Aristotelianism; but nonetheless, it was at bottom Peripatetic,

and it had to be mastered—or at any rate, ﬁrst approached—within the

context of the Peripatetic philosophy. It needed explaining as soon as a

student opened the Predications—and where better to look for the elements

of an explanation than in the essay itself?

So what—according to the Predications or Categories—are predicates and

subjects? An ancient commentator on the essay would ﬁrst tell you that

Aristotle took the term ‘κατηγορεῖν’ from the law-courts, where it meant

‘to accuse (someone of something)’ or ‘to impute (something to someone)’.

The imputation went in the accusative, the imputee in the genitive. So an

indignant Demosthenes refers to an enemy as

he who, o earth and gods, imputes philippism to me …

(cor 294)⁴⁰

The imputation or predicate is philippism, the imputee or subject is Demo-

sthenes.

Next, the ancient commentator would remind you that there are in

principle three possible answers to the question ‘What are predicates?’: ‘They

are words’; ‘They are thoughts’; ‘They are things.’ He would explain how

each answer was confronted by insuperable objections—and then he would

pull a rather tatty rabbit from his hat: ‘Predicates are words insofar as they

are used to signify things by way of thoughts.’ So each of the three possible

answers is partially correct.

⁴⁰ ὃς γὰρ ἐμοῦ φιλιππισμόν, ὦ γῆ καὶ θεοί, κατηγορεῖ ...

116 Predicates and Subjects

The rabbit was generally fathered on Porphyry. Here is a passage from

Dexippus’ set of questions and answers on the Categories—a work which is

largely cribbed from Porphyry:

The predicates are not the entities themselves but the expressions which signify the

thoughts and the objects. When they say

Animal is predicated of man

they say that the expression signiﬁcant of animal—which is the name animal—is

predicated of the thought signiﬁed by the expression man and of the object which is

subject to that thought; for being predicated is a property of signiﬁcant utterances

which signify thoughts and objects.

(in Cat 10.25–32)⁴¹

Animals are not predicated, nor are thoughts of animals or the concept of an

animal: what is predicated is the word ‘animal’. But not every word can be

predicated: not the word ‘blityri’, for example, because though it is a word, it

has no sense; not the words ‘often’ or ‘and’ or ‘through’, for example, because

although they are words endowed with senses, they do not signify anything.

(What does that mean? An answer will emerge in a later context.)

The question ‘What sort of item is a predicate?’, and hence Dexippus’

answer to it, were developed in terms of a sketchy semantic theory which

ancient scholars discovered in the ﬁrst few lines of Aristotle’s de Interpretatione.

According to that theory, the word ‘animal’ stands for or means, primarily,

the thought or concept of an animal—or perhaps the thought or concept

of an animal which is in the soul of the man from whose mouth the word

‘animal’ emerges. That thought or concept fortunately resembles animals or

an animal. And so the word ‘animal’ signiﬁes, derivatively, animals.

But is not that theory quite ridiculous? And must not any account of

predication which is based upon it be hooted or booted off the philosophical

stage? Yes; and No. For although the Porphyrean account of predication

was conceived and presented in terms of a certain set of semantic ideas,

and although those ideas—whether or not they are Aristotelian—are indeed

laughably inept, nonetheless the account of predication is in fact independent

of the ideas. After all, the account amounts (so far) to no more than this:

⁴¹ οὐκ αὐτὰ οὖν τὰ ὄντα αἱ κατηγορίαι, ἀλλ᾿ αἱ σημαίνουσαι λέξεις τὰ νοήματα καὶ

τὰ πράγματα. ὅταν γὰρ λέγωσι τὸ ζῷον κατὰ τοῦ ἀνθρώπου κατηγορεῖται, λέγουσιν ὅτι ἡ

σημαντικὴ λέξις τοῦ ζῴου, ἥτις ἐστὶ τὸ ζῷον ὄνομα, κατὰ τοῦ σημαινομένου νοήματος ὑπὸ τῆς

ἄνθρωπος λέξεως καὶ τοῦ ὑποκειμένου τούτῳ πράγματος κατηγορεῖται· τὸ γὰρ κατηγορεῖσθαι

τῶν σημαντικῶν φωνῶν ἦν ἴδιον, αἳ σημαίνουσι τὰ νοήματα καὶ τὰ πράγματα.

Porphyrean Predicates 117

predicates are a sort of signiﬁcant expression. To that you may attach whatever

theory of signiﬁcation pleases you—or no theory at all.

However that may be, the Porphyrean account of predication was gen-

erally—if not universally—accepted in late antiquity. And of course it

must have commended itself to anyone who thought that, at least in the

simplest cases, the predicate of a sentence is its verb; for verbs are signiﬁcant

expressions. But if the Porphyrean account triumphed in antiquity, not many

modern Aristotelians will be found to support it. After all, it goes ﬂatly against

numerous texts in which Aristotle clearly takes predicates to be something

other than expressions.

Here are a couple of familiar examples, the ﬁrst from the de Interpretatione

and the second from the Prior Analytics:

Of objects, some are universal and some particular—by universal I mean such as can

be predicated of several items and by particular such as cannot (for example, man is

a universal and Callias a particular).

(Int 17a38–b1)⁴²

Objects, or πράγματα, come in two sizes, and what makes the difference is

their predicability: that is to say, here it is objects and not expressions which

are supposed to be predicated. Or again:

Of all the things which there are, some are such that they are not predicated truly

and universally of anything else (e.g. Cleon, Callias—whatever is individual and

perceptible) but other items are predicated of them (each of the two, after all, is

a man and an animal). Other things are themselves predicated of other items and

yet nothing else is earlier predicated of them. Yet further things both are themselves

predicated of other items and have other items predicated of them (e.g. man of

Callias and animal of man).

(APr 43a25–32)⁴³

Aristotle is classifying ‘the things which there are’—entities or τὰ ὄντα:itis

among existent things, and not among expressions which apply to existent

things, that he locates predicates.

⁴² ἐπεὶ δέ ἐστι τὰ μὲν καθόλου τῶν πραγμάτων τὰ δὲ καθ ᾿ ἕκαστον—λέγω δὲ καθόλου

μὲν ὃ ἐπὶ πλειόνων πέφυκε κατηγορεῖσθαι, καθ ᾿ ἕκαστον δὲ ὃ μή, οἷον ἄνθρωπος μὲν τῶν

καθόλου Καλλίας δὲ τῶν καθ ᾿ ἕκαστον.

⁴³ ἁπάντων δὴ τῶν ὄντων τὰ μέν ἐστι τοιαῦτα ὥστε κατὰ μηδενὸς ἄλλου κατηγορεῖσθαι

ἀληθῶς καθόλου (οἷον Κλέων καὶ Καλλίας καὶ τὸ καθ ᾿ ἕκαστον καὶ αἰσθητόν), κατὰ δὲ

τούτων ἄλλα (καὶ γὰρ ἄνθρωπος καὶ ζῷον ἑκάτερος τούτων ἐστί)· τὰ δ᾿ αὐτὰ μὲν κατ᾿ ἄλλων

κατηγορεῖται, κατὰ δὲ τούτων ἄλλα πρότερον οὐ κατηγορεῖται· τὰ δὲ καὶ αὐτὰ ἄλλων καὶ

αὐτῶν ἕτερα (οἷον ἄνθρωπος Καλλίου καὶ ἀνθρώπου ζῷον).

118 Predicates and Subjects

So the Porphyrean account of predication, whatever its philosophical

credentials, is not a faithful account of Aristotelian predication.

Before I answer, on Porphyry’s behalf, that seemingly deﬁnitive objection,

there is another and less daunting obstacle which needs to be confronted.

Dexippussaysthatin

Animal is predicated of man

the predicate ‘is predicated of the thought signiﬁed by the expression man and

of the object which is subject to that thought’. There the phrase ‘the object

which is subject to that thought’ gives the Greek ‘τὸ ὑποκείμενον τούτῳ

πρᾶγμα’—and that looks for all the world as though it is just a longwinded

way of saying ‘the subject’. In any event, whereas Dexippus insists that

predicates are expressions, the items of which they are predicated are said to

be not expressions but thoughts or objects. There are any number of ancient

texts in which it is taken for granted that you predicate things of objects and

not of expressions, and hence that subjects are objects and not names of objects

(except, of course, in the special case in which the object is itself a name).

And after all, the subject of a sentence is surely what the sentence is

about, and sentences are generally about things rather than about their

names:

Charlie is me darling

says something about a prince, not about his Christian name. True, you can

talk about signiﬁcant expressions, for example:

Constantinople is a very long word

is about the name of a city and not about a city. But just as it is not Charlie

but the name ‘Charlie’ which is a part of the sentence

Charlie is me darling,

in the same way—or so the logicians assure us—it is not the name of the

city but a name of that name which features in the sentence

Constantinople is a very long word.

And apart from a few oddities—such as

‘Sentence’ is the last word in this sentence

—sentences are not about their own parts.

So, if Dexippus is right, then on the Porphyrean account of predication,

subjects are one sort of thing and predicates another. But in that case it

cannot be a true account of Aristotelian predication: it breaks on the fact that

Aristotelian subjects and Aristotelian predicates are homogenous: any item

which is a predicate may also be a subject, any item which is a subject may

also be a predicate.

Porphyrean Predicates 119

That is indeed an obstacle to Porphyry’s account—or to the Dexippan

version of the account; but it is not an insuperable obstacle. For there are

several ways of surmounting it. You might, for example, construe a term

not as an expression but rather as an ordered pair consisting of a signiﬁcant

expression and its extension.⁴⁴ Then if x is <‘A’, A> and y is <‘B’, B>,

xispredicatedofyifandonlyif‘A’holdsgoodofB.Needlesstosay,

there is nothing like that in any ancient text. Rather, an ancient Porphyrean

would maintain that, just as predicates are signiﬁcant expressions, so too are

subjects. Such a view is present in all those authors who claim that, at least

in the simplest cases, it is names which are the subjects of sentences. And

there are several ancient texts in which it is explicitly stated that the subject of

this or that proposition is an expression. No doubt it is, so to speak, natural

to think of a subject—especially if it is called a ὑποκείμενον or underlying

item—as if it were something external to the sentence in which it receives a

predicate; and the ancients frequently fell into that natural way of thinking.

But it was a fall: their theory virtually required them to take subjects to be

expressions—and when they thought about the matter, they did.

In that case, does it not follow that all predicative sentences are about

signiﬁcant expressions? And is that not absurd? It is absurd; but it does not

follow. What follows is that a subject of a sentence is not an item about which

the sentence has something to say: at best, it is a signiﬁcant expression which

indicates an item about which the sentence has something to say.

But if the requirement of homogeneity may be met by insisting that

subjects as well as predicates are signiﬁcant expressions, what about the ﬁrst

of the two objections to Porphyry’s account—the objection that it doesn’t

ﬁt the Aristotelian texts? It must now appear even more formidable; for if

it is hard to accommodate linguistic predicates to the texts, surely it is even

harder to accommodate linguistic subjects? When, for example, Aristotle

distinguishes in the Categories between items which are said of a subject

and items which are in a subject, the subjects are not expressions. White,

he says, is in a subject—namely, a body. He does not mean that the word

‘white’ is in the word ‘body’, but that the colour white is in that bag of

bones.

Porphyry was aware of the objection. The very posing of the question

‘What sort of item is an Aristotelian predicate?’ suggests that the matter is not

immediately clear; and Porphyry knew that the various candidate answers

⁴⁴ This possibility was brought to my attention by Otto Bruun.