[lit-ideas] Re: Language isn't arbitrary, convention based after all

  • From: Donal McEvoy <donalmcevoyuk@xxxxxxxxxxx>
  • To: "lit-ideas@xxxxxxxxxxxxx" <lit-ideas@xxxxxxxxxxxxx>
  • Date: Tue, 29 Oct 2013 14:42:55 +0000 (GMT)

The results obtained are perhaps not so surprising. After
all, our “emotional” reaction to the sound of a nail scraping a stone may
differ from the “emotional” reaction to the sound of a smooth-running stream:
these are non-linguistic sounds with different “emotional” ‘affects and 
so we might not be so surprised that the units of linguistic sounds [phonemes]
may have different “emotional” affects and effects [independent of their
semantics and the tone of their speaker]. 
This leaves the question of what kind of explanation is
needed for these different ‘affects and effects’. This post indicates how an
approach in terms of World 3, World 2 and World 1 content might help frame the
questions, but also how the answers may vary depending on the combination of W3,
W2 and W1 elements involved. There is also a problem in how we might devise
tests as to which elements play which role in the explanation: for example,
even if it were virtually universal that a scraping nail was an irritating
sound compared to a smooth-running stream, it would be left open whether (and
to what extent) the explanation lay in the W1 physics of the relevant sounds,
the W2 of their mental ‘reception’, or the W3 of their ‘cultural’ “meaning”? It
might not be easy or straightforward to disentangle the W1-W2-W3 aspects so as
to produce a testable theory as to their respective roles (or even a
well-developed untestable theory). There also may not be any detailed
explanation in a specific case [e.g.
scraping nail] that holds so that it is equally valid across the whole field of
A while ago I posted on Popper’s views of artistic and
mathematical World 3 content:- suggesting that in the case of certain artistic
content, in contrast to mathematical content, that W3 content may be co-variant
or interdependent with its World 1 expression in a way that the relation
between W3 and W1 content is not merely ‘arbitrary’ or ‘conventional’. This is
connected perhaps with the issue raised by the authors – that certain kinds of
“emotional meaning” may be ‘intrinsic’ to certain W1 “formants”. Likewise we
might say certain kinds of “emotional meaning” or artistic affect are 
to certain W1 “formants” in music; likewise with the “formants” of scraping
nails and running streams.
For certain purposes there are meanings of “meaning” where
we might say that W3 “meaning” is (partially) determined by the W1 vehicle by
which that W3 content is conveyed – where by “(partially) determined” we mean
that the W1 vehicle is not merely arbitrarily related to the W3 content it
conveys. But a comparable issue arises in relation to W1 sound where that sound
is considered without reference to any W3 content – for example, the sound of a
scraping nail or the sound of a phoneme when considered merely as W1 entities.
Even such merely W1 entities may have ‘effects and affects’ at a W2 and a W3
level. It is doubtful that because they are merely W1 entities therefore these 
and affects’ may be understood in purely W1 terms.
These meanings of “meaning” may be connected with “emotional
meaning” rather than mere ‘propositional content’. They are perhaps connected
with the kinds of meaning that typically concern the artist rather than the
scientist. The artist’s work may not be about merely ‘propositional content’
devoid of “emotional meaning” (and considered merely for its truth or falsity) 
about conveying ‘content’ that is allied with appropriate ‘form’ for its
expression. (Thus achievement in terms of “appropriate form” helps explain why
we may regard an artistic work as a towering achievement even if we take its
‘propositional content’ as largely false, and may regard an artistic work as
poor even if we take its ‘propositional content’ as largely true [e.g. it seems 
to ‘propose’ that requited
love makes people happy].)
There are large, involved issues here; any analysis should
be careful not to run away with itself. As indicated, it is not even easy to
always keep clear the distinction between W3 and W1 content, never mind clearly
address their possible inter-dependence and the role of the intermediate W2. A
question raised in my previous posts surfaces again here: even where W3 content
is inter-dependent on its W1 vehicle (so their relation is not merely
‘arbitrary’, such that another W1 vehicle would do just as well) are we to
explain this interdependence in terms of it being an affect of W1, an affect of
W2 ‘reception’, or as nevertheless a W3 affect – or all these (perhaps
Here this post gives me an opportunity to correct something
important, where my previous post was not sufficiently qualified. 
Previously I wrote:-
“We might also emphasize that when we speak we make physical
sound – but the physics of sound does not constitute the meaningful content of
what we say: and no World 1 level of explanation is adequate to explain the
meaningful content of what we say.” 
Now this is arguably false (the argument hinges on how we
understand “adequate”). It certainly needs to be qualified. Corrected by
qualification, it should read:-
 “…no World 1 level of
explanation is fully adequate to
For there are W1 levels of explanation that are adequate up to a point to help 
explain meaningful
content and to differentiate it from meaningless content: for example, we can
explain differences between musical and non-musical sounds in purely W1 terms
and this explanation may be ‘adequate’ up
to a point for distinguishing music from non-musical sounds (though highly
incomplete from the POV of the W3 understanding of music). Similarly there may
be W1 levels of explanation that are adequate up to a point to help explain 
differences between linguistic sounds
and non-linguistic sounds: the ‘phoneme’, considered as a purely W1 entity or
unit, may itself be part of such a W1 explanation. 
In the case of music, the gist of the W1 explanation (i.e. the explanation a 
physicist might
give) that distinguishes musical and non-musical sounds is that musical sounds 
are organised patterns in terms of their
physical frequencies in a way that non-musical sounds are not. But when we try
to understand further the character of their organization and their musical
affects and effects, we eventually move beyond issues that are merely W1
issues. Principles of physics may explain why the sounds of the opening of
Beethoven’s Fifth are different – in
their “formants” – to the sounds of tables being overturned and glasses
smashing, and may give a principled explanation to differentiate musical from
non-musical sounds: but there are not principles of physics fully adequate to 
explain how and why
the Fifth’s opening differs in its
artistic content from the opening of Beethoven’s Ninth.
Using this opportunity, there is more in my previous posts that
requires more careful formulation. That post continued:-
“The physics of sound may be essential to speech but it is
epiphenomenal to its World 3 content: and this is obvious because we may, by
convention [such as changing the natural language we speak], alter the World 1
encodement of the language without altering its World 3 content – so that “The
snow is white” and “Die Schnee ist weiss” may differ in World 1 terms but not
in terms of their World 3 content.”
This should be corrected to:- “The physics of sound may be
essential to speech but it may be epiphenomenal to its World 3 content…” – for 
there may be examples where the
physics of sound may be more than
epiphenomenal to its W3 content. (The researchers may be concerned with
some such examples).
In the overall context of my posts, these lapses may not
have been very misleading – for clearly the posts sought to distinguish cases
[like “mathematical content”] where W1 may be merely “epiphenomenal” to W3
content, from cases [like “artistic content”] where it may not. Nevertheless,
it is easy enough to write in a way that clouds the issues by not being careful
enough; and my posts’ failings illustrate this.
Despite these lapses, the final words of the relevant paragraph
may still hold:-
“World 1 provides instruments or vehicles for embodying and
conveying World 3 content but that World 3 content is distinct from any merely
World 1 level of the instrument or vehicle that conveys it – and, at least in
cases like mathematics and in the case of propositions, the World 3 content may
be invariant throughout differing World 1 forms of its expression.” 
What needs to be emphasised is the “may be”, for this “may be”
only for certain cases.
What the researchers may be touching on are cases where
certain aspects of “content” may not be invariant throughout differing W1 forms 
of its expression [as per “artistic content” and as per “emotional meaning”]. 
But we should
beware thinking that the explanation for this W1-W3 ‘co-variance’ is
straightforward – for example, it is not clear how we would test or otherwise
show whether the ‘co-variance’ is an ‘affect’ of  W1, or W2, or W3 –  or an 
inter-linked W1-W2-W3 ‘affect’ [q: were we a different kind of being (say
in terms of our World 3, W2 ‘reception’, or our W1 brains) would we experience
certain W1 sounds differently in terms of their “emotional meaning”; might such
a different being generally find the stream makes an irritating sound and the
scraping nail makes a relaxing sound]? It is not even clear that such
‘co-variance’ is always of some uniform type or that it is not always a varying
product of a complex interplay of factors.
There is also a general question left open: does the
research severely undermine the view
that the physical form of linguistic expression is largely only ‘arbitrarily’ 
or ‘conventionally’ related to its
“meaning” or ‘content’? Even if the research is pointing up something
interesting in terms of “emotional meaning”, it may leave largely untouched
other kinds of meaning – in particular, the “meaning” or content we might term
‘propositional content’. For ‘propositional content’ may be very largely 
by the W1 linguistic means used to convey it – so that such ‘propositional
content’ may be very largely unaffected, for many research purposes, by the W1
form of its expression. Even the most diligent and scrupulous scientist
considering the authors’ paper may conclude that they will not lose much of
importance if they focus on developing the ‘propositional content’ of their
theories as best they can, without worrying much about the phonemes they use.   

These difficulties are aside from the fact that our reaction
to specific phonemes in terms of their “emotional quality”, in cases where 
those phonemes lack a semantics or tone, may be coloured by our acquaintance 
other phonemes in cases where the “emotional quality” of those other phonemes 
is affected by their semantics and tone [i.e. the study may suffer from fatal 
flaws of possible ‘cross-contamination’ of its
linguistic samples in this way]. 
But the more fatal flaw is that the study is oblivious to
fundamental distinctions of a W1-W2-W3 type.


On Tuesday, 29 October 2013, 3:33, Torgeir Fjeld <torgeir_fjeld@xxxxxxxx> wrote:
"A bit of numerological mysticism can do wonders for the academic branch known 
as Linguistic Patricide, innit?" phatic muttered.

It was a dark and gloomy night at the No Holds Barred cafe in uptown Florida.

"You've been reading the Digest again, 'aven't ya," responded Beanieman 

phatic nodded, hesitantly. 

"Why it's apt - psycho-acoustically speaking - that Darth Vader wasn't called 
Barth Vaber"

The relationship between the meaning of a word and the letter strings of which 
it is comprised is usually thought to be arbitrary. That is, the meaning of a 
word is dictated by convention and the emotional tone of the speaker. Strip 
these away and the sounds of the letter groupings themselves - known as 
phonemes - are generally considered meaningless. At least that's been a popular 
view for some time.

But now a study has been published that challenges this account. Blake 
Myers-Schulz and his colleagues show that the shift in sound from some phonemes 
to others carries emotional meaning of its own, quite independent from word 
meanings or tone of voice.

Human speech creates sound at different frequencies. Myers-Schulz and his team 
focused on the changes in certain frequency peaks in speech - known as formants 
- as nonsense words were spoken. Specifically, they divided nonsense words into 
those in which the first two formants went from low to high (e.g. bupaba, 
pafabi, mipaba) and those in which this sound shift was reversed, going high to 
low (e.g. dugada, tatoku, gadigu). They were matched on many other sound 
features, such as plosives, nasality, intonation and volume.

Thirty-two adult participants were shown pairs of these nonsense words on a 
computer screen, one of which always went low to high, the other high to low 
(in terms of formant shifts). Together with the words, two pictures were shown, 
one positive, one negative (e.g. a cute puppy and a snarling dog). The 
participants' job was to allocate the two nonsense words to the two pictures in 
the way that seemed most appropriate. The key finding was that 80 per cent of 
the time, they matched the word that had the low-high sound shift with the 
positive picture and the high-to-low word with the negative picture.

It was a similar story when 20 more adult participants performed the same task 
but with the words spoken by a computer programme rather than shown visually. 
In this case, they matched the low-to-high nonsense words with the positive 
pictures on 65 per cent of occasions - still far more often than you'd expect 
based on chance alone.

The findings suggest that strings of phonemes (the sounds that comprise words) 
have an emotional quality of their own, quite separate from any word meaning or 
the tone or volume of an utterance. This emotional meaning is conveyed purely 
by the acoustic properties of the word as the sound frequencies change from one 
phoneme to the next.

There could be intriguing real-life applications for this research in terms of 
marketing and PR because the implication is that some words convey positive 
emotion simply by virtue of their acoustic properties, above and beyond any 
literal word meaning. "Even in artistic contexts, such as film and literature, 
these acoustic principles could be applied to evoke a particular emotional 
subtext," the researchers said. "Indeed our data suggest that 'Darth Vader' is 
an acoustically more appropriate name for an intergalactic miscreant than 
'Barth Vaber'."

Myers-Schulz B, Pujara M, Wolf RC and Koenigs M (2013). Inherent emotional 
quality of human speech sounds. Cognition and emotion, 27 (6), 1105-13 PMID: 
Mvh / Yours,

Torgeir Fjeld
Gdansk, Poland

Blogs: http://phatic.blogspot.com // http://norsketegn.blogspot.com
Web: http://independent.academia.edu/TorgeirFjeld

Other related posts: