My best part of the day, is dropping my 5-year old to school. Ah! The questions and insights...
Of late, she has become reticent and prefers to listen to Music. Pretty soon, she got bored of her kids' songs. So, I function as her DJ. When a new song plays, she either commands, "I don't like it. Could you change daddy" or she remains quiet.
I noticed that, whenever "Jack and Diane" plays, she sings along. So, I brought in a collection of John Mellencamp. Bumner! She did not seem to like anything other than "Jack and Diane".
So, I kept trying other singers in that Genre.
For her age, her taste is pretty nuanced. She would not like a slow song. Yet she particularly likes Micheal Buble. The traditional taxonomy of musical genres is not very useful for me, to select songs for her.
Pandora did not work for me,beyond a point, partly due to the short rides. Since I am a music lover myself, I curate the songs for her and of late, she rarely says, "I don't like it".
I can almost externalize my daughter's taste pattern, into a new custom genre, which I would call, the 'dida-5'. Her taste is not going to be the same, 2 years from now. It would become 'dida-7'.I am pretty sure, there would be others whose taste fall into 'dida-5'.
The trajectory of her taste could be reliably predicted and would emerge like a fractal, with data from other consumers, bound together, in a connected consumption platform.
I would call this, 'computer-generated' custom taxonomy as 'fractonomy'. I used taste as an anchor for classification. However, Taste could be easily replaced by any other attribute, in this "Everything is Miscellaneous" Era.
I like the term 'fractonomy' and what it opens up.
ReplyDeleteMusic categories are meta-stable: artists can belong to one group for years but when they become influential they create their own and destabilise cognates/contiguous/related categories.
The trajectory of some taste-choices is path-dependent. If you like jazz, you are very very likely to end up digging on Coltrane.
Tiny differences in early conditions can lead to massive changes later on.
Good DJs can be defined by their mastery of the fractonomy.
Observing the fractonomy is possible but manipulating it not - there's a taxonomic Uncertainty Principle thing going on.
I'd be curious to see how 'taste' mapped onto other attributes (eg function, social environment, introversion/extrovert etc)
Thanks. Nice language game.
Thanks for your comments and tweeting the skinner link.
ReplyDeleteMany folks have questions about this. So, I thought I could offer some explanation.
Taxonomy:
A taxonomy is simply a classification of things.
Two Characteristics:
1. Give exact name to everything.
2. How is this thing related to other things?
This is hard and might not be very usable for a non-expert to 'use' or 'consume' a system as they should understand the context well and know the terms used.
Folksonomy:
Give a bunch of names that you think is relevant.
Don't worry about relationships and stuff like that.
But, you might call it cat and someone feline and there is no recognition that these are synonyms.
There is no semantic inference that, a car is a vehicle.
There is an absence of context. There is no way to deduce that, iparty and "pump it up" are adjacent categories, in a child birthday context. So, if you are searching for iparty, how can you get "pump it up".
Fractonomy solves these problems:
It can introduce rigor to folksonomy-based systems, by recognizing synonyms, making semantic inferences and creating a "dynamic category", which could be named as well.
It is dynamic in the sense, an item could potentially be transported to another category, if new attributes are added or existing ones are changed.