OCKHAM ON CONNOTATION

Index

Here is a new translation of chapter 10 of Book I of William of Ockham's Summa Logicae, where he defines 'connotative' and 'absolute' terms. It provides an interesting insight into the original meaning of these terms. The modern distinction has its direct ancestry in Mill, though he claimed it was derived from the scholastic philosophers. The passage here shows that Mill's definition of connotation was closer to the definition given by scholastic logicians like Ockham than some of his detractors have claimed.

Ockham's definition hangs upon the distinction between what the later scholastic philosophers (and, after that, early modern philosophers such as Locke) called the real essence of a thing (the quid rei, the 'what' of the thing) and the nominal essence (the quid nominis, the 'what' of the name). A connotative term, says Ockham, is one whose meaning can be expressed by a nominal definition - a form of words expressing the meaning of the name. A purely absolute term is one which does not, properly speaking, have a nominal definition.

The idea of a nominal definition is difficult. For more about that, see here. Ockham gives two explanations, as follows.

1. A term has a nominal definition when it is necessary to put one part of the definition in the nominative case and another in an oblique case. For example, the term 'something white' (album) has the nominal definition 'something having whiteness' (aliquid habens albedinem) in which one word (something – aliquid) is in the nominative and another (albedinem) in an oblique case. Ockham's point seems to be that nominative signifies the subject of the sentence, so 'a white thing' signifies the actual thing that is white. But an oblique case such as the accusative or genitive or other case signifies something to which the subject bears some relation (other than identity), and so cannot be the actual thing that is white. Thus 'something just', 'something human' are connotative, because they are equivalent to 'something having justice', 'something white', 'something having humanity', whereas 'man', 'animal', 'goat' are not, because they are equivalent to expressions of the form 'something having – ness'.

2. A term has a nominal definition when it can be explained by different expressions whose meaning is different, in that one part of one expression signifies something that is not conveyed in the same way by some part of the other expression. Because of this, says Ockham, none of them is properly a nominal definition. For example, one person explains what the name 'angel' means by saying 'I understand by an angel a substance abstracted from matter', another saying 'an angel is an intellectual and incorruptible substance', and another 'an angel is a simple substance that does not join with any other'. (The example of angels is one that Locke also uses. See here).

Neither is entirely clear. According to the first definition, 'man', 'animal', 'goat', 'stone', 'tree' are not connotative because we cannot defined them in terms of expressions of the form 'something having – ness'. But can't we say something like 'something having man-ness', or 'something informed by goat-hood'? This appears to rely on an accident of lexicology. And in any case we do have words like 'humanity', 'asineity'. Ockham's argument relies here on a claim he makes elsewhere (ibid. chapter 5) that names like 'cowship', 'goathood', and so on are never or rarely found, and that this is because the concrete term 'man' is in any case synonymous with the corresponding abstract term 'humanity', so for 'animal' and 'animality'.

According to the second definition, a name is only connotative when there is a form of words that uniquely specifies what the name means. Otherwise it is absolute. Is Ockham suggesting that non-connotative terms have more than one meaning? Or that none of the differently meaning explanations (as with the case of 'angel') give the true meaning of the word? From what he says here, it is not altogether clear.

Perhaps he is suggesting that the meaning of absolute terms derives from our unmediated grasp of the objects they signify? Thus absolute terms are all truly predicable of (past or present) things. See here. By contrast, there can be connotative terms for all kinds of things that have never existed, even for things which could not possibly exist, such as the chimaera - see here. Spade argues for such an interpretation here. As Spade argues, if any connotative term can be replaced by its nominal definition, it would follow that any nominal definition can be expanded into a definition that consists only of absolute terms. This seems consistent with Ockham's program of analysing away spurious entities by his theory of connotation. (Including the Aristotelian "categories", for example).

I have consistently translated 'quid nominis' as nominal essence, and 'definitio exprimens quid nominis' as 'nominal definition', in order to express the connection between the scholastic and early modern philosophy. Other translators use the more literal 'expression explaining what the name means'.

The passage suggests that Mill's definition of connotation was not so far from the mark as some of his critics have claimed. To be sure, Mill said that 'man' connotes being a man, whereas Ockham holds that 'man' is not connotative. But Mill held that all general names can be defined, and so all are connotative, in Ockham's sense. Mill, like Locke, though that to be a man, or of the species man, and to have a right to the name man, is the same thing. Ockham, by contrast, does not hold that all general names have a definition, so not all general names are connotative, in Mill's sense.

Note also that while Mill held that all general terms are connotative, he thought that all proper names and demonstratives are absolute, in Ockham's sense, because they cannot be explicated by a form of speech. (To give an example parallel to Ockham's example about the angel, we can talk about 'the teacher of Alexander', the pupil of Socrates and so on, and everyone knows what you mean. But these are not definitions of 'Aristotle'). See here for Mill's definition of connotation.



LatinEnglish
Latin English
[CAP. 10. DE DIVISIONE NOMINUM IN MERE AESOLUTA ET CONNOTATIVA] C. 10. ON THE DIVISION OF NAMES INTO PURELY ABSOLUTE AND CONNOTATIVE
Postquam de nominibus concretis et abstractis est discussum, nunc de alia divisione nominum, quibus scholastici frequenter utuntur, est dicendum. Unde sciendum quod nominum quaedam sunt absoluta mere, quaedam sunt connotativa. Nomina mere absoluta sunt illa quae non significant aliquid principaliter et aliud vel idem secundario, sed quidquid significatur per illud nomen, aeque primo significatur, sicut patet de hoc nomine 'animal' quod non significat nisi boves, asinos et homines, et sic de aliis animalibus, et non significat unum primo et aliud secundario, ita quod oporteat aliquid significari in recto et aliud in obliquo, nec in definitione exprimente quid nominis oportet ponere talia distincta in diversis casibus vel aliquod verbum adiectivum. After concrete and abstract names have been discussed, now we have to speak about another division of names which scholastics frequently use. Wherefore, you should know that certain names are 'purely absolute', certain are 'connotative'. Purely absolute names are those which do not signify something principally and something else (or the same thing) secondarily, but rather, whatever is signified by that name, is equally signified primarily, thus the name 'animal' clearly does not signify anything but cattle, donkeys and men, and so for other animals, and does not signify one primarily and another secondarily in such a way that something has to be signified in the nominative case and another in an oblique case, and in the nominal definition it is not necessary to put such distinct terms in different cases, or to use some participle.
Immo, proprie loquendo talia nomina non habent definitionem exprimentem quid nominis, quia proprie loquendo unius nominis habentis definitionem exprimentem quid nominis est una definitio explicans quid nominis, sic scilicet quod talis nominis non sunt diversae orationes exprimentes quid nominis habentes partes distinctas, quarum aliqua significat aliquid quod non eodem modo importatur per aliquam partem alterius orationis. Indeed, properly speaking, such names do not have a nominal definition, since, properly speaking, there is one definition of a name that has a nominal definition, explicating the nominal essence, thus it is evident that of such a name there are not different nominal definitions having distinct parts, of which one signifies something that is not conveyed in the same way by some part of the other expression.
Sed talia quantum ad quid nominis possunt aliquo modo pluribus orationibus non easdem res secundum suas partes significantibus explicari, et ideo nulla earum est proprie definitio exprimens quid nominis. But, as far as their nominal definition is concerned, such names can be explained in some way by several expressions that do not signify the same things by their parts. And on that account none of them is properly a nominal definition.
Verbi gratia 'angelus' est nomen mere absolutum, saltem si non sit nomen officii sed tantum substantiae. Et istius nominis non est aliqua una definitio exprimens quid nominis, nam unus explicat quid hoc nomen significat, sic dicendo 'intelligo per angelum substantiam abstractam a materia'; alius sic 'angelus est substantia intellectualis et incorruptibilis'; alius sic 'angelus est substantia simplex, non componens cum alio'. For example, 'angel' is a purely absolute name (at least if it is not the name of a function, but of a substance only). And there is not one nominal definition of this name, for one person explains what the name means by saying 'I understand by an angel a substance abstracted from matter', another saying 'an angel is an intellectual and incorruptible substance', and another 'an angel is a simple substance that does not join with any other'.
Et ita bene explicat unus quid significat hoc nomen sicut alius. Et tamen aliquis terminus positus in una oratione significat aliquid quod non significatur eodem modo per terminum alterius orationis, et ideo nulla earum est proprie definitio exprimens quid nominis. And so one explains what the name means just as well as the other. And yet some term put in one expression signifies something that is not signified in the same way by some term in the other expression, and for that reason none of them is properly a nominal definition.
Et ita est de nominibus mere absolutis quod stricte loquendo nullum eorum habet definitionem exprimentem quid nominis. Talia autem nomina sunt huiusmodi 'homo', 'animal', 'capra', 'Iapis', 'arbor', 'ignis', 'terra, 'aqua', 'caelum', 'albedo', 'nigredo', 'calor', 'dulcedo', 'odor', 'sapor' et huiusmodi. And so it is for purely absolute names that (strictly speaking) none of them has a nominal definition. Now such names are as follows: 'man', 'animal', 'goat', 'stone', 'tree', 'fire', 'earth', 'water', 'heaven', 'whiteness', 'blackness', 'heat', 'sweetness', 'smell', 'taste', and the like.
Nomen autem connotativum est illud quod significat aliquid primario et aliquid secundario. Et tale nomen proprie habet definitionem exprimentem quid nominis, et frequenter oportet ponere unum illius definitionis in recto et aliud in obliquo. Sicut est de hoc nomine 'album, nam 'album' habet definitionem exprimentem quid nominis, in qua una dictio ponitur in recto et alia in obliquo. But a connotative name is one that signifies something primarily and something secondarily. And such a name does properly have a nominal definition, and often it is necessary to put one part of that definition in the nominative case and another in an oblique case. So it is for the name 'white', for 'white' has a nominal definition, in which one word is put in the nominative and another one in an oblique case.
Unde si quaeras, quid significat hoc nomen 'album', dices quod illud idem quod ista oratio tota 'aliquid informatum albedine' vel 'aliquid habens albedinem'. Et patet quod una pars orationis istius ponitur in recto et alia in obliquo. Wherefore, if you ask what the name 'a white thing' signifies, you will say: that name signifies the same as the entire expression 'something informed by whiteness' or 'something having whiteness'. And it is clear that one part of the expression is put in the nominative and another part in an oblique case.
Potest etiam aliquando aliquod verbum cadere in definitione exprimente quid nominis, sicut si quaeras, quid significat hoc nomen 'causa', potest dici quod idem quod haec oratio 'aliquid ad cuius esse sequitur aliud' vel 'aliquid potens producere aliud', vel aliquid huiusmodi. Sometimes too a verb can occur in the nominal definition, so, for instance, if you ask what the name 'cause' signifies, it can be said, the same as 'something on whose existence something else is a consequence' or 'something that can produce another thing', or something of that sort.
Huiusmodi autem nomina connotativa sunt omnia nomina concreta primo modo dicta, de quibus dictum est in quinto capitulo. Et hoc quia talia concreta significant unum in recto et aliud in obliquo; hoc est dictu, in definitione exprimente quid nominis debet poni unus rectus, significans unam rem, et alius obliquus, significans aliam rem, sicut patet de omnibus talibus 'iustus', 'albus', 'animatus', 'humanum', et sic de aliis. Now such connotative names include all concrete names of the first sort (which were mentioned in c. 5). And this is because such concrete names signify one thing in the nominative and another in an oblique case; that is to say, in the nominal definition there ought to be put one nominative term, signifying one thing, and another oblique term, signifying another thing, as is clear for all such names as 'just', 'white', 'animate', 'human', and so on.
Huiusmodi etiam nomina sunt omnia nomina relativa, quia semper in sua definitione ponuntur diversa idem diversis modis, vel distincta, significantia, sicut patet de hoc nomine 'simile'. Such names also include all relative names, since in their definition there are always put different terms signifying the same thing in different ways or distinct things, such as is is clear for the name 'similar'.
Si enim definiatur simile, debet dici sic 'simile est aliquid habens qualitatem talem qualem habet aliud', vel aliquo modo consimili debet definiri. Unde de exemplis non est magna cura. For if 'similar' is defined, it should be said thus: 'the similar is something having a quality such as another thing has', or it ought to be defined in some similar way. Wherefore there is no great importance attached to these examples.
Ex quo patet quod hoc commune 'nomen en connotativum' est superius ad hoc commune 'nomen relativum', et hoc accipiendo hoc commune 'nomen connotativum' largissime. From which it is clear that the common term 'connotative name' is of a higher genus than the common term 'relative name', taking here the common term 'connotative name' in the broadest sense.
Talia etiam nomina sunt omnia nomina pertinentia ad genus quantitatis, secundum illos qui ponunt quantitatem non esse aliam rem a substantia et qualitate, sicut 'corpus', secundum eos, debet poni nomen connotativum. Such names also include all names pertaining to the genus of quantity, according to those who propose that quantity is not another thing from substance and quality, so that 'body', according to them, should be held to be a connotative name.
Unde secundum eos debet dici quod corpus non est aliud nisi aliqua res habens partem distantem a parte secundum longum, latum et profundum; et quantitas continua et permanens non est nisi res habens partem distantem a parte, ita quod ista est definitio exprimens quid nominis ipsius. Wherefore, according to them, it ought to be said that a body is nothing but some thing having part distant from part in respect of length, breadth and depth; and continuous and permanent quantity is nothing but some thing having part distant from part', so that this is a nominal definition.
Tales etiam consequenter habent ponere quod 'figura', 'curvitas', 'rectitudo', 'longitudo', 'latitudo' et huiusmodi sunt nomina connotativa. Immo, qui ponunt quod quaelibet res est substantia vel qualitas, habent ponere quod omnia contenta in aliis praedicamentis a substantia et qualitate sunt nomina connotativa; et etiam quaedam de genere qualitatis sunt connotativa, sicut ostendetur inferius. Such persons also have to maintain that 'figure', 'curvedness', 'rightness', 'length', 'breadth' and the like are connotative names. Indeed, those who propose that every thing is either a substance or a quality have to suppose that all the contents in categories other than substance and quality are connotative names; and even certain names in the category of quality are connotative, as will be shown below.
Sub istis etiam nominibus comprehenduntur omnia talia 'verum', 'bonum', 'unum', 'potentia', 'actus', 'intellectus', 'intelligibile', 'voluntas', 'volibile' et huiusmodi. Unde de intellectu est sciendum quod habet quid nominis istud 'intellectus est anima potens intelligere', ita quod anima significatur per rectum et actus intelligendi per aliam partem. Under these names are also comprehended all such as 'true', 'good', 'one', 'power', 'act', 'intellect', 'intelligible', 'will', 'willible'and the like. Wherefore, concerning intellect you should know that it has this nominal essence: 'Intellect is soul able to understand.', so that the soul is signified by the nominative, and the act of understanding by the other part.
Hoc autem nomen 'intelligibile' est nomen connotativum, et significat intellectum tam in recto quam in obliquo, quia definitio sua est ista 'intelligibile est aliquid apprehensibile ab intellectu'. Ibi intellectus significatur per hoc nomen 'aliquid', et per istum obliquum 'intellectu' significatur etiam intellectus. But the name 'intelligible' is a connotative name, and signifies the intellect in the nominative just as in an oblique case, because its definition is 'the intelligible thing is something that can be apprehended by the intellect.' Here the intellect is signified by the name 'something', and the intellect is also signified by the oblique 'by an intellect'.
Et eodem modo dicendum est de 'vero' et 'bono', quia 'verum', quod ponitur convertibile cum 'ente', significat idem quod 'intelligibile'. 'Bonum' etiam, quod est convertibiIe cum 'ente', significat idem quod haec oratio 'aliquid secundum rectam rationem volibile vel diligibile'. And we should say the same of 'true' and 'good', since 'true', which is supposed convertible with 'being', signifies the same as 'intelligible'. 'Good' too, which is convertible with 'being',signifies the same as the expression 'something willable or lovable according to right reason'.







THE LOGIC MUSEUM Copyright (introduction and translation) © E.D.Buckner 2006.