What’s new

Three articles came together for me this week.

  1. New Scientist magazine reported on an ambitious project to use artificial intelligence (AI) to speak with animals.
  2. New Scientist also reported on the astounding amount of code in old programming languages still present in financial processing software.
  3. A friend who writes and teaches about Celtic paganism and horse goddesses posted a link to a 2017 article arguing that Joseph Campbell got a lot wrong in his work on mythology.

What does it mean?

The first article states “AI is good at language,” and describes a current project to decode the sounds made by sperm whales using AI to look for categories in massive databases. The technique “is great at detecting patterns and can neatly sort whale calls, say, into piles based on their acoustic properties, but often can’t tell you what those piles relate to.” The researchers realize they need to correlate the whale’s clicking vocalizations, called codas, with whale behavior and conclude “This is still a long way from deciphering meaning.” The article falls, I fear, into a long, long pattern of hype about AI, with the headline “Dr Dolittle machines: How AI is helping us talk to the animals” belied by the actual content of the article.

Among the many social problems made even more visible by the corona virus pandemic, attempts to provide financial aid to people found that some computer systems were overwhelmed “with the flood of people applying for welfare benefits – and hardly anyone around knew how to fix things,” as stated in the second article. These welfare processing systems, it turns out, still have massive amounts of code written in COBOL and other old programming languages. Ageing programmers, like the COBOL Cowboys, were called into service to try to fix the situation. But it often turns out that you don’t just need an experienced COBOL programmer, you need the COBOL programmer who wrote the original code. “Opaque turns of phrase, plus coding conventions that can vary significantly between domains or even organisations, make deciphering a specific bit of software difficult for an outsider.”

While Campbell’s views shifted across his many works, he is best known for arguing for the similarity of myths across many cultures. His thesis is summarized in the title of one his books, The Hero with a Thousand Faces, in which, according to the Joseph Campbell Foundation, “Campbell formulated the dual schemas of the Hero’s Journey, a universal motif of adventure and transformation that runs through all of humanity’s mythic traditions, and of the Cosmogonic Cycle, the stories of world-creation and -dissolution that have marked cultures around the world and across the centuries.” To the contrary, the 2017 article by Jeana Jorgensen quotes Sara Cleto as writing: “By saying `all these stories are the same,’ we lose what stories mean in different contexts and, especially, what they can mean to people that come from cultures that are not our own.”

What does it mean for you?

Humans look for and see patterns, perhaps due to evolution, since pattern recognition can help us survive. The theme I see in these three articles is that problems can arise through overuse of categories or patterns. It is more than just the difference between lumpers and splitters in taxonomy, it is more than whether patterns or categories help or hinder our thinking, the issue is our overuse of the general at the expense of the particular and our failure to respect the meaning an individual puts on an experience.

I can give more examples.

  • What some cite as Freud’s belief in universal symbols in dreams is better described as the need to interpret symbols in the individual’s context.
  • New Scientist also recently reported on efforts to use AI to identify emotions from facial expressions, but Lisa Feldman Barrett is quoted as questioning the fundamentals of the approach because “the use of various expressions varies noticeably between cultures.”
  • Again, from New Scientist, the Y2K bug lingers because the specific local ways in which the fix was implemented have caused other problems.

Local customs and cultures matter. Individuals matter.

In my interactions with people, I am constantly on the watch for indications that my interpretation of a situation is not the same as the other people I am working with. I can’t assume that the pattern, the categories, or the meaning I see is the same as what others see. I am listening for the implicit statement “That’s not how I see it.”

In my specific area of creating probability models, one person may say to another “wait, what is the sample space you are using?” which means that we are not thinking about the situation in the same way. We need to back up and reexamine our different interpretations of the situation.

In our quest for shared meaning, in our quest to establish an organizational culture, and in our quest to communicate clearly, we can use the amazing human skill of creating categories and of looking for patterns, but we must always be carefully alert for and we must always respect the particular, the local, and the individual.  

Where can you learn more?

In support of the recognition of the particular over the general, I urge you to watch your life over the next days and recognize those situations where another person interprets an event differently than you do. I think of them as Rashomon moments. For me, Facebook is a rich source of amazing and sometimes shocking cases where consensus is disrupted by someone who just views the whole situation in a different light. Wow, I say to myself.

I am adamant about using images that are in the public domain and about giving credit to the creator of the image, but I was stymied in this case by my inability to find an open use image of the sled Rosebud from the movie Citizen Kane, a powerful image, I think, of the particular nature of symbols. I substituted a rosebud photo. Did you get the reference? Or was it too particular to me?

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.