Serendip is an independent site partnering with faculty at multiple colleges and universities around the world. Happy exploring!

The Classification Problem: Implications for Intentionality





The Classification Problem



Taxonomists and systematists are interested in understanding evolutionary relationships between living organisms.  This in turn is motivated by an interest to understand the forces and factors that influence evolution within taxa and in so doing, evolution in general. 

Mycology, the study of fungi, is an interesting test case for the systematist because to infer evolutionary process the researcher must first come to some consensus on the pattern from which one is inferring that process.  Species is the unit of measure that biologist use to categorize the patterns of living diversity, but fungi display a unique set characteristics that make it very difficult to delineate species.  There are many versions of the species concept and fungi are problematic for all of them.  In other words, how does a researcher infer process (evolution) from pattern (species), when the organisms defy classification - when fungi do not fall into a clear pattern. This presents a philosophical consideration I call the Classification Problem that has implications for epistemology in general, the nature of reality and constructs of it. What follows in this hyperlink is a brief overview of relevant fungal characteristics to set up the Classification Problem. 



There are now three interesting intersection with the Classification Problem (CP) and evolving systems - particularly with the storytelling, intentional agents:


  • First, do categories exist independent of the classifier? What is the relationship between Nature vs Classifier; Code vs Decoder; Reality vs Perceiver and perhaps Maker vs Audience.  


  • Second, how does one classify or even identify intentionality as posed by Anne with regards to Paul’s six (non)intentional images: "Can we distinguish between non-intentional and intentional artifacts?"  And if the answer to the first question leads us to conclude that there is no construction of reality without a perceiver (in physics: all possibilities collapsing into one because of an interaction/observer) then Tim’s beaver dam vs blindspot is a resounding “Blindspot”. In the parlance of classification; intentionality is only a construct of a classifier that believes they themselves are intentional.  Intentionality is an artifact of a unique and particular sieve that filters reality. We identify intentionality if we believe there is a maker behind the artifact - as we ourselves are makers.


  • Finally, if intentionality is reduced to the assumption of a “maker”, then what happens if we let go of the assumption of a “maker”?  What becomes of human intentionality if the universe and all that is in it has no beginning, no end, no maker, only the interaction between code and decoder? Does it even matter whether we assume there is an intentional maker or not, if there is no why to distinguish between the two alternatives?




Let’s start to peel this three layered onion by trying to define classification.


  • the process of delineating. 
  • the process of discriminating variation (information) into discrete non-overlapping units.




Does this definition suffice? Does all, or for that matter, any variation, fall into discrete, non-overlapping units?  Or is the universe continuous and overlapping? To pose the question slightly differently, What’s in a name? 




…a rose by any other name would smell as sweet.  - Shakespeare's Romeo and Juliet, 1594.


Juliet’s lament cuts to the basic assumption that categories are “real”.


By “real” I mean that discrete, non-overlapping units exist independent of the observer.


Therefore, we and any model builder perceive them because they exist.

But discrete, non-overlapping units could be what’s left of “reality” after we filter it… or the only parts we can perceive due to the limits of our perception. There may not be any discrete, non-overlapping units at all? All variation/information may be continuous and overlapping or alternatively, composed of random units - singularity some would argue is functionally the same as pure randomness. 


Illustrations of the discrete vs continuous dilemma and implications for inferring Evolutionary.

Diverging Populations, Two phylogenetic scenarios for nervous tissue, Punctuated vs Gradualism, Patterns of Biodiversity as a snapshot of Evolutionary process


Can we distinguish between these alternatives?


To explore theses alternatives more closely and perhaps begin to imagine others, I’d like to present two demonstrations that frame the discovery versus filter dichotomy.  That is to say, assuming there is a "reality", is it discovered or constructed?


Essentially this is the problem of, “Does a tree make a sound when it falls in the woods?”


I would argue that something is there whether or not there is a “perceiver” or “decoder” present to receive the information/code.  However, sound is a construction dependant on the type of decoder. The type of decoder has a particular filter by which to sieve reality.


To flesh out the decoder component of this problem let’s examine human perception briefly.


The Searching-Perception Problem: Some information is imperceptable due to lack of a proper search template.




The Discrimination-Perception Problem: Other information cannot be decoded because of a structural blind spot due to the limits and constraints of a particular model-builder.


Index to Stereograms



What does this set of observations demonstrate:


  • We don’t know what we don’t know
  • Some things we don’t know we can come to know
  • Other things we don’t know we cannot know
  • With the aid of technology (human constructs) we can enlarge the set of things we can know



An interesting set of questions arise from these observations:


  • What are the limits to our knowing?
  • Is the set of things we don’t know infinite or finite?
  • If it is finite we might be making progress.  If it is infinite, then progress has no meaning at least in the normal use of the term. And if we are not making progress, what are we doing?


Ended here... Start with summary comments:



To summarize, I posit that Nature/Code/Reality is not knowable/meaningful without a Classifier/Decoder/Perceiver.  Or... Fish is Fish?


"Physical concepts are free creations of the human mind, and are not, however it may seem, uniquely determined by the external world”

(Einstein & Infeld, 1938).




"So What?" as Peter asked.  "Of course, observers are biased."

Indeed! But it follows then that reality is constructed by biased agents and therefore, one never knows what is "real".  Furhtermore, when one storytelling agent interacts with another, neither are sure of the others biases and ability to faithfully decode each others intentions/meanings?


So,... what of our intentions?



For the time being let’s not worry about reality.  Yes, I am using the terminology of a reality – an out there.  But, operationally speaking, I propose an agnostic stance.  We can have the same exact conversation if we want to talk in completely relativistic terms of constructions.  My point is the same. There is either only constructs or there are only constructs of something we don’t have direct access to.  



Back to the point, a reality dependant on a perceiver/decoder suggests a simple solution to the intentionality problem highlighted by Anne’s provocative question, “How does one know something can be attributed to intentionality.”  Re: Paul’s six intentional images.


One attributes intentionality to phenomena/artifacts that are believed by an observer to have a maker.

We see ourselves as makers/creators, thus often attribute intention to other phenomena/artifacts.


The obvious question then becomes how does one determine if a phenomena/artifact had a maker?  I mean to distinguish a maker from a causal process, but the only way I can do that is to be caught in a tautology.  A maker has  intention, while a causal process does not.  

Perhaps this is a case of Gödel's paradox: a formal system can be either internally consistent, but incomplete or the system can be complete, but inconsistent.  For sake of completing the argument, allow me to be inconsistent.


Back to the question of distinguishing between a phenomena that has a maker from one that does not? I think this question may be unanswerable.  But whether or not one attributes a phenomena to maker, does have implications for the meaning of intentionality that I would like to examine.


If there is a maker of the universe, then all interactions, phenomena and artifacts are intentional.  But what if there is no beginning, no end, no maker? What if there is only interactions between agents/entities collapsing possibilities that in turn affect new possibilities in an ever evolving open system? Or alternatively all agents are “makers”, making and re-making the universe as they interact. Or, there is a maker, but the intention is unknowable, therefore operationally these alternatives are synonymous.  


If there is no maker, there is no intention.  If there is no intention, there is only interactions between agents.  Some of the interactions generate new possibilities: some resonant, some dissonant.

But what about "storytelling agents"? Or, as Peter asked, "Why teach, why communicate at all?"

This still leaves me wondering about the perception I have about myself as a conceiver of counterfactuals, an agent of free will and free choice.  Do I really have free will?  I obviously effect change via my interactions, but can I direct my effects to a particular goal? 


And here we are back at the same question that Paul and Anne has left us with, and so I will, too:


  • Can an agent affect change with a particular/intentional goal in mind and what are the limits to that effect?


I don’t have an answer, but I would like offer a few lines of inquiry:

What is the difference between a tree or plant  model builder/decoder and a human decoder in their ability to affect change? 


Is it even meaningful to ask whether or not a plant affects change with a goal in mind?  But then, how does one explain a sunflower plant tracking the sun all day from east to west, then at night turning back to the east in anticipation of the sunrise?  See the following time-lapse video of this sunflower behavior.


Is there a useful distinction between plant and human intentionality?  Between model builders and storytelling model builders?  I would like to pose a new classification: simple intentionality and counterfactual intentionality?  Let me explain.



Plants and other life forms have evolved interactive feedback loops between the environment and the individual.  If individuals vary, ie, there are many variants - populations of these feedback loops, then those feedback loops that result in greater longevity for those variants come to persist.  That is to say, in so far as the feedback loops account for information from and about the environment that results in behavior that increases longevity/persistence/Darwinian fitness, the feedback loops are said to model the universe.



The difference between human interactive feedback loops and that of a plant is that human interactive feedback loops not only feedback to the environment, but in addition feedback to counterfactuals about the environment.  Therefore, plants are environmental model builders (possess simple intentionality) and humans are that and additionally, counterfactual model builders (display counterfactual intentionality). If this counterfactual model building results in longevity then this type of model building persist. 


One important difference between counterfactual model building over simple environmental model building is the iteration unit is drastically, dramatically and profoundly reduced.  Where a plant must wait generations and needs many variant types of model builders to ever modify the environmental feedback loops, counterfactual model builders have feedback loops within an individual (as opposed to a population) and does not need to wait a generation to modify the model.  In fact, counterfactual models can almost instantly be adjusted.  This reduction in the iteration unit between code and decoder, I suspect, has profound effects on evolution, the rate and the course. It also, suggest some limitations of counterfactual model building.  Counterfactual models can, and I would argue are, more efemeral and less robust as they are within an individual, not a populations.  It is much more likely for an individual to slip into oblivion than an entire population, to say nothing of whether or not there is a genetic component that even accounts for some or all to the variability in counterfactual models.  Environmental models are preserved in DNA (at least to some extent), only the ability to produce counterfactual models are in DNA. The counterfactual model itself is only in the individual or text or product that the individual created and thus could be said to be housed in culture derived from a populations of individuals.  It might and could be debated as to what is more robust over time, text and other artifacts of counterfactual intentionality or DNA, but my bet is on DNA. 


I would like to end with some Taoist reflections on intentionality. Here are a few quotes attributed to the counterfactual model builder named, Lao Tsu:


The task is accomplished,
but the Sage doesn't seek credit or take pride in the accomplishment. 

I’ve read another translation of this sentiment put this way:

 The master creates for creation sake, without intention or purpose.



The world is a sacred vessel.
It should not be meddled with.
It should not be owned.

If you try to meddle with it, you will ruin it.
If you try to own it, you will lose it.





A student gains each day.
A follower of the Way (Tao) loses something each day.
Loss after loss until arriving at Non Action (Wu Wei).* [non-intentionality?]





Post new comment

The content of this field is kept private and will not be shown publicly.
To prevent automated spam submissions leave this field empty.
1 + 1 =
Solve this simple math problem and enter the result. E.g. for 1+3, enter 4.