0:12
Welcome back to the MOOC on Nexus and Sustainability!
In this lesson, we will present how
complex system thinking can be used to handle quantitative analysis,
avoiding the pitfalls of reductionism we illustrated earlier.
In the first session here,
we will discuss the characteristics of a "grammar".
What is a grammar and why it is useful?
Of course, a grammar is a familiar concept in natural language.
So in order to show our grammar can be used to
organize information even quantitative information,
we start with an example based on the natural language.
So this is a letter,
a letter of rejection of
an applicant that made an application to Harvard University and has been rejected.
So we read "Dear applicant,
We appreciate your interest in Harvard University,
We would like to start off that this year,
we had a lot of very fancy applicants.
Unfortunately, as you could probably tell from
the thickness of the envelope. You were not accepted.
If it's worth any consolation,
we waited until the last minute.
I'm sure that the college, so on, so on,
so on... We wish you good luck.
The Office of Admission."
Why is it important to waste all this time on this letter?
It is because this letter has been written by a computer.
It has not been written by a real person.
So, how a computer can know how to write a letter?
So then, where the grammar enters into play, in reality,
the grammar is establish a set of relation between the whole,
the letter of rejection,
the scope to tell the person that has been rejected in a nice way.
Then it has a functional element: greetings, introduction,
bragging (you get a lot of applicants), contrast,
dropping the bomb (you were not accepted and so on).
Then, this functional element can be expressed by using different structural elements.
For instance, dropping the bomb here could be: we turning down your application,
you were not accepted,
your application was not approved, you suck,
all these structural elements are fulfilling the goal of this functional elements.
Then, there are different levels of analysis.
These here represented an equivalent class of structural element at the level n-2,
whatever you could hear, n-1,
will be functional to the level n,
but then you will have additional complication.
You can have, in bragging,
among those who applied this year,
we get 38 HOTSHOT of type 1, 275 of HOTSHOT of type 2,
and the Hotshot could be many scorers: class presidents, child prodigies, doesn't matter.
You can make it more and more complex.
As long as you keep
a set of expected relations over the different elements that have to do different things.
So we could get into different types
of structural element on the right and whatever structural element here,
you combine according to the rules,
you will always generate a letter of rejection.
For example, this is an old letter of rejection is
different from the previous one because it's using different structural element,
but still it does the same result.
As long as you have the right combination of
lower level structural element and the right choice of structural elements here,
before was: you were not accepted,
you suck, your application didn't make it.
Again, the same structure of the letter,
and then because of the structure that the grammar gives to it,
you can keep using the different elements,
structural or functional, in an effective way.
So the concept of grammar can be associated
with definition of relation between two non-equivalent sets.
Functional elements, they are elements that puts together,
managed to express a property.
So, if you put the list of greetings, intro, bragging, contrast,
you are generating a rejection letter;
and then the structural elements,
they have an organization determined by the ontological characteristic.
We can imagine as a string of characters or word,
that these are mapping into a message.
So what you see here,
in a way, is what we saw before as holons,
in the sense, this is the structural element and this is the functional element,
together do the job.
Why is this is important?
Because then in quantitative terms,
you can define equivalent class.
One functional type, this would be the human beings,
so this is the heart of the human being.
But let's imagine that you take out the heart, still
you will have what is called in jargon, a natural niche.
What is expected by the context about the heart;
the quantity and quality of blood supply, requirements of inputs.
These expected functional types can be expressed by two different structural types.
We already saw this when we discussed of the holon.
So basically, what we are dealing with
the representation of a complex system or a metabolic pattern,
we can use a definition of functional types
to individuate what are the aspects of the system we want to
study and then depending on
the structural times that we are considering how we are measuring them,
we can generate quantitative analysis.
Let's give an example how to do an analysis of food.
We already discussed this about the problem with the energy at the food accounting.
So you have a category of final consumption,
how much food will be eaten,
then would be the losses,
how much food is required by the agricultural sector of production,
and this is the gross requirement of food that has to be produced
to consume internally, and then, of course,
you have a domestic production either you
produce this food on your own or you have to import.
So this functional characteristics of the system are relevant for different questions.
I mean this is to know whether you have
enough food for guaranteed good quality of intake and this is the internal overhead;
how much this is about technology and choice of production?
How much the production is consuming internally?
This is what you need,
then this is how much land,
water and other natural resources you need locally to have a domestic production.
And this is how much you are dependent on imports (externalization).
These are all characteristics that we are defining in semantic terms,
there are still no numbers in it.
Then we could start getting numbers in, if you remember,
this is the amount of kilos required for in the diet,
but then we saw that this amount of kilos is not
equal to the amount of kilos that you need in production.
This is much more and why is that?
Because a large part is, especially,
in the developed countries of kilos go in producing
a feed or alcohol that is not directly getting into the diet.
So the kilos of grain required to
produce animal products is consumed by the United States but indirectly.
We will discuss this again.
This is not the place where to explain this number.
What is relevant here, is that by using
a taxonomy of a country we can have numbers they are referring to semantic category.
Why is this important?
Because let's imagine, if we are looking at a food system in
relation to the diet of the population,
we can put the numbers in it.
As a matter of fact, we are getting a little more
sophisticated rather than kilos of grain.
When talking about nutrition,
we can have a vector in which we can put carbohydrates,
proteins, and fat; and then we will have, still,
the same semantic structure about import, domestic supply,
how much you need a gross of vegetable grains or animal product,
and that how much goes directly in the final consumption,
how much is consumed by the agricultural sector.
You see we have the same structure of variable
before but this time we are measuring carbohydrate,
petajoule of carbohydrate, protein, and fats.
You see, and this is how much of this energy goes in the household.
This is the total and then we can split in a vector on which
we can have a carbohydrate, protein, and fat.
But we could use exactly the same grammar
to look at agricultural problems: Do we have enough land?
Do we have enough water?
Then in this case,
we have a different story about food.
It is about the agricultural side.
So we can use the same grammar,
import, domestic supply, final consumption,
but this time, instead of using protein, carbohydrate and fat and all of it,
we shall put in vegetable grains, oil, and animals.
So we have different numbers, here, you see.
The numbers here, as difference, we have
a different quantitative analysis but is reflecting,
describing the same system,
and then giving the same answer.
Of course, we can put here money,
we can put here whatever.
We have the same system defined in semantic term,
in terms of functional elements,
and then depending on what type of structural element we are mapping,
we will get different numbers.
I am going quick here because we will give a lot of examples of this later on,
so we don't want to waste time in getting into the particular number.
Mario GiampietroICREA Research Professor
Andrea SaltelliGuest researcher
Tarik SerranoPost-Doc Researcher
