Dienstag, 13. April 2010

Knowing - metaphorically spoken

How do languages express abstract concepts, such as to know something?

Let us have a look at the means the speakers of some Indo-European languages have chosen long ago, how they used to say "I know":
In the ancient Indo-European languages, there was the verb *weid- 'to behold'. In more than one language of the family the present perfect form of the verb, which can be reconstructed as *woida, originally meaning 'I have beheld', acquired the meaning 'I know'. In the course of the centuries the old perfect form became an independent verb with a meaning different from that of its origin.
The development seems quite transparent and even plausible - something I have perceived/beheld is something I have got to 'know'!

In the old Indian language, we have veda - 'I know' (well known from the most ancient religious texts of the Indian people - the Veda, as in Rig Veda, Atharva Veda...). The ancient Greek language obviously passed the same process: Greek οίδα ('I know' - exactly matching Old Indian veda), the perfect form derived from the before mentioned Proto-Indo-European root, also means 'I know'.
However, in the Latin language the process observed before has either not taken place, or its result has been lost. Latin
vīdī merely means 'I have seen'.

As to Germanic languages such as Gothic and Old High German, Gothic wait , witum and the OHG verb weizen (the NHG 'ich weiß') - are of the same origin and meaning as the Indian, Greek, and Latin forms.

Our observations suggest that a parallel semantic development has taken place:
from a verb with the meaning (roughly) 'to see' via its perfect form 'to have seen' to a new verb with the meaning 'to know'. This has been the case in several Indo-European branches such as Indo-Aryan, Greek, and Germanic. As we deal with one Proto-Indo-European root
*woida , the process might already have taken place in the proto-language.

Abstractly speaking, we find that the above mentioned languages express the abstract state of knowing sth. by metaphorically employing the sensual concept of seeing.

Latin, like other languages, has developed various means to denote the concept of knowing. We find that the word -scire 'to know' has originated (retrograde) from the negative ne-scio 'I don't know' (note that not the positive -scio is the original word, but the negative ne-scio!).
-scio in ne-scio is a form of lat. secui, 'I have cut', the perfect of seco, secare, 'to cut, to part'.
So the Latin language associates the abstract concepts of knowledge and understanding with processes of dissecting, of cutting into parts, of discerning the parts of a matter - with analysis (btw., the greek word analysis has an etymological background similar to our "cutting" words).
In modern German, we have a parallel concept for discerning: the verb unterscheiden. The simplex scheiden originally means 'to divide', 'to part', 'to cut'.

The English equivalent to unterscheiden is the verb discern
from Latin discernere (from dis- 'off', 'away' + cernere 'distinguish', 'separate', 'sift'). It refers to the same notion - both in English and in Latin!

Up to now, we have found two quite divergent metaphorical concepts of
knowing: to see refers to a mental concept which is partly associated with the notion of wholeness, whereas to discern links our acquisition of knowledge to acts of dividing, of cutting apart - of analysis. Two mental concepts - of synthesis and of analysis - they are supposed to be complementary in the process of acquiring knowledge.

A third way Indo-European languages denote
knowing can be shown in the etymology of English to know. It originates from PIE *gneh3- , 'to know', which became Gothic kunnan, Old Norse kna, Old English cnawan, OHG knaen: Interestingly, the Old English word seems to have been mainly used to express 'to know from experience'; it also meant 'to be able'. A cognate of this word is the German können.
The Anglo-Saxons used two distinct words for the concept of knowing, witan (see wit; cf. our discussion of *woida above) and cnawan. The latter seems to refer to a more practical point of view, as in the Modern English phrase to know better "to have learned from experience" (first attested 1704). The Modern English Know-how "technical expertise"was first recorded 1838 in American English. So in Anglo-Saxon we have two verbs with somehow complementary shades of meaning, witan used in mainly theoretical contexts, and cnawan for knowledge with practical implications. There was another Old English preterite-present verb absorbing a third sense of "to know," that of "to know how to do something" (in addition to "to know as a fact" and "to be acquainted with" something or someone). Its original perfect participle, couth, has survived only in its negation uncouth. But note that could and can are from the same root; another cognate word derived from this origin is cunning.

There are further interesting metaphorical concepts for "knowing", e. g. to grasp - German be-greifen : we get to know something by touching it, by taking possession of it (compare German "Besitz ergreifen").

I tried to demonstrate by reference to a few examples how languages can express the mental concept of "to know something" -
by means of metaphors from the world of our senses -
with verbs that originally meant "to see", "to cut", "to grasp", ...

The examples should also hint at the historical dimension of metaphorical speech - they were taken from the very confined field of ancient Indo-European languages and their modern descendants.


  1. Another English word for knowledge is of course "ken" -- both for "extent of knowledge" (as in "beyond our ken" but also "to know, be acquainted with, recognise" ("do you ken John Peel?"), with Germanic cognates.

  2. Thanks for pointing to this very interesting word. Similar to the German word "kennen", it seems to be a causative form of the word "cunnan". The original meaning of Old English "cennan" probably was "to make to know" = "explain". :-)