Refining early Basque criteria

Larry Trask larryt at
Tue Dec 21 10:45:18 UTC 1999

[ moderator re-formatted ]

Pat Ryan writes:

[on Mama-papa words like Basque <ama> 'mother']


> I feel that you may be mixing apples with oranges here.

> I would, myself, be quite sceptical of any claim that an imitative word like
> <miau> indicated anything more than an attempt to capture the quintessential
> acoustic impression of a cat-call.

> But, 1) there is nothing "imitative" about <ama> for 'mother';

Agreed, but I didn't mean to say there was.  The point is not that such words
are imitative; the point is that they are *motivated*.

> 2) more
> importantly, <ama> does not have the form we would expect from childish
> babbling, which, I hope you would agree, would be along the lines of
> C(1)V(1)C(1)V(1).

Yes; babbling is stereotypically of the form CVCV (reduplicated).  But not
invariably so.

> I gladly concede that <mama>, <d/tad/ta>, <kaka>, etc. are childish attempts
> to render other words, e.g. <*?ama> and <*?atV>, etc.

Oh, no -- this is not the point at all.  See below.

> but there is nothing
> that I know which *necessitates* or universally *inclines* children all over
> the world to connect /m/ with 'motherhood' or /d/ or /t/ with
> 'paternity' --- short of some universalistic sound-symbolism argument, which
> I provisionally do not accept.

No; this is a misunderstanding.

Children do not make any such connections as those suggested at all.
The point is that nursery words are *not* invented by children: they are
invented by adults.

Once children reach the stage at which they are beginning to have enough
control over their vocal tracts to produce speech sounds consistently, they
behave in a highly consistent fashion, as argued by Jakobson as long ago as
1941. The first vowel they learn to produce is [a] -- the easiest vowel to
produce, since it requires minimal tongue action.  The first consonants they
produce are labials -- [m], [b], [p] -- presumably because these require no
tongue control.  The next consonants they learn are coronals -- [n], [d], [t]
-- presumably because these require no more than the raising of the tip/blade
of the tongue.  Velars, which require bending of the tongue, come later, as do
other consonants.

Accordingly, the first consistent noises the eager parents hear from the child
are things like [(m)ama], [(b)aba], [(p)apa], followed by [(t)ata], [(d)ada],
and so on.  It is at this point that the delighted parents decide that their
child is trying to speak -- which is very doubtful -- and assume happily that
the little bugger is trying to say 'mother' and 'father'.  The assumption that
the kid is trying to say 'mother', rather than 'tickle' or 'telephone' or
'banana' is one made *entirely* by the parents.  Jumping to this conclusion,
the happy parents begin to speak back to the child, using what they fondly --
but wrongly -- believe to be the child's own words.  In this way, such
"mama-papa words" -- as we call them -- can become institutionalized in adult

So, such words recur in lots of languages as a direct consequence of the
observable universal progression of speech-sound production in infants, coupled
with the widespread tendency of parents to interpret these early sounds as
having specific meanings.  That's all.

> I would suggest rather that these very ancient words have been retained in a
> substantially unchanged form because of the strong emotional significance
> they have in most human societies.

But this is fanciful, and there exists a far simpler explanation.  Just listen
to an infant producing its first speech sounds, and you have your explanation.
Nothing more elaborate is called for.

Larry Trask
University of Sussex
Brighton BN1 9QH

larryt at

More information about the Indo-european mailing list