Written by Ian Sharp PhD on April 14th, 2021
You are listening to The Good Doctor Sharp on DoctorsInTech.com
and today I'm talkin about talkin... and all I can say is... "We got it wrong".
Current language and the interpretation of communication is flawed and
I think it's in need of an upgrade.
Language today is imprecise, inefficient, open to interpretation and subject to miscommunication.
And at what price?
In a world... where digital echo-chambers close communication loops
at unprecedented rates and at scale, we've seen where that leads, and we can do better.
Over the past few decades, we've seen
the advent of more technology developed to
try and take current language as input, and have it interpreted by
software (think, Siri, Alexa, Facebook forum analytics).
I think we are going in the wrong direction, not because of the technology strategies,
but rather because
the quality of the output of any software can only be as good
as the quality of the information passed in.
The quality of language, as used today is poorly crafted, and not as usable as it could be.
For these reasons, I therefore propose that
we first improve the quality of language by making it more precise.
In general, I propose that human speech would be more effectively
modeled after pure mathematics,
rather than how it is modeled today. It's time for a deliberate design.
How can we design an upgrade to human language to have the highest impact on our conversation efficiency?
Let's first define "efficiency" in terms of information theory.
Secondly, let's select a conversation classification that is quite common.
Information theory maintains that conversations are no more than an exchange of information.
The classification of conversation I've chosen is one in
which the information exchanged is limited to:
"beliefs about outcome distributions"
I've chosen this classification of conversation because it is
omnipresent. This conversation is inherent in every decision making process.
This kind of conversation is had both in face-to-face meetings and in digital land.
For this reason, upgrading just this one conversation
I think is going to have the largest positive impact with the highest frequency across
the highest number of people.
As a matter of fact, most of my consults fall into this conversation category,
as do many group internet forums, and social media discussion in general.
In this type of conversation, each person exchanges information about what they expect
the future distribution of outcomes to look like.
They don't know that's what they're doing... but that's what they're doing.
As such, the boundaries on their belief distributions are often unclear.
Furthermore, the conversation (information exchange) often
drowns on, longer than necessary, and often spirals away
into chasms where at times, little to no information of value is exchanged at all.
I'm going to give a "for example", now and make this real to you.
Imagine you are in pursuite of truth, and are having a conversation
with a contentious conversationalist. They are
in pursuit of nothing more than making themselves appear "right" by making you appear "not 100% right".
To them, if you are in any way wrong, then they are right.
And it goes like this.
You
"Americans are overweight"
contentious conversationalist
"No, they're not! My cousin's uncle is skinny as a rail, and he's American."
The implicit point being
Look! I got you! You're expected outcome for Americans and being overweight is wrong.
By extension, I must be really smart and in fact smarter than you.
If ALL Americans are not overweight, then you're wrong. He's smug and
happily sneering at you right now. Look at this goofy face (in real life, or on an internet forum)
We've all have had the unfortunate situation, to be part of a conversation like this.
You almost just want to roll your eyes right? Is it even a waste of time to respond?
Couldn't we all know I didn't mean "Every American in the world is XYZ" -
Do I actually have to say it? Well, technically the weakness of human speech does
imply that I must actually be crazy to believe 100% of Americans are overweight right?
Since that's what I "said".
How pathetic is language? Now if you go back and change your point, you're
going to appear unsure of yourself.
You
"OF COURSE NOT ALL...!
I just meant MOST Americans are overweight"
The contentious conversationalist may not even be listening, they're happy they
got you to back-step, or side peddle along a point that is obvious to begin with.
Anyways, what a waste of time. And you know you're not done yet either, because
what is "Most" anyway. So let's try this again.
"Americans are overweight : 1 sigma" (σ meaning 1 standard deviation)
contentious conversationalist
"No, they're not! My cousin's uncle is skinny as a rail, and he's American."
You
"CDC says 73% of Americans are overweight.
Americans are overweight: 1 Sigma. That's covered by 1σ.
No one's talking about your 2σ cousin's uncle.. just you..
Do you have any additional information to add to this conversation?”
The contentious conversationalist DOES actually understand distributions on some level,
but they only understand one kind of distribution... the binary. Like
"CAPITAL"
who thinks "either its broke or it works".
Because they can only comprehend binary things... they're actually very upset... almost constantly.
Most of my consults,
at a high-level, boil down to someone telling me their expectation of a distribution of future outcomes,
then I respond and exchange information about my expected distribution of outcomes, and whether or not
our beliefs about "outcome distributions" are the same or not.
To be honest I'd much rather start injecting probability distribution charts into
my conversations.
"Your software will be ready in 6 months : 1 sigma σ"
But back to my question earlier about "At what price?"
Mental health professionals in recent decades
suspect binary thinking, (i.e "all or nothing thinking"),
may be implicated in cluster personality disorders (Beck, Freeman, Davis, & Associates 2004).
If it's possible that a limited language could encourage "all or nothing thinking",
then we might expect the outcome of limited language use in a rapid-rate, closed-loop digital-media-land,
to more quickly percolate cluster personality disorder, if we were to adopt a diathesis-stress model.
Bottom line 1: Rationalizing about outcomes with all or nothing thinking -
it's not a good way to teach people to think.
Bottom line 2: Language that promotes binary thinking - it's not a good language to rely on.
Future work
To further discuss advocacy to update human speech, to make it more efficient, less error-prone, and increase the information exchange rate, get in touch. Mental health professionals, are you already trying to upgrade your clients thinking? Would like to utilize a platform to conduct clinical research and are interested in this topic? If so, then sign up today.