Like this article? We recommend
Like this article? We recommend
Like this article? We recommend
Installing a new map of reality (paradigm shift)
Although it is usually not the role of the Generative Analysis to install a new map of reality for someone, there are exceptions. For example, software engineering is often subject to paradigm shifts. If we consider the old paradigm to be the old map, then a true paradigm shift destroys the old map and replaces it with a new one. A case in point is the shift from procedural programming to object-oriented programming that occurred in the 1990's. This was a very uncomfortable time for those developers who were heavily invested in the old paradigm and suddenly had adopt and reskill in the new paradigm. We are seeing a similar thing today, where Generative AI is radically changing the paradigm for software development. Because of this we may need to install some new maps of reality, especially for developers that might be invested in some of the more cultic software engineering approaches based on hand crafting, loyalty and groupthink. The process is simple in principle, but often very difficult in practice because emotional investment in the old map can lead to strong resistance. Here is how we would do it:
- Use Why questions to take someone up to the edge of their existing map.
- Ask a Why question that can't be answered using the old map (is disallowed) but can be answered in the new map. This is the point at which there may be a lot of resistance.
- Introduce the new map by answering the disallowed Why question by using the new map.
Consider the following example where A is the Generative Analyst and D is a developer:
A: I hear you are a bit sceptical about using UML in development
D: Yeah, you just don't need it – it's a waste of time
A: Why do you think it is a waste of time?
D: It doesn't give us anything. It's best just to write the code.
A: Why is it best to just write the code?
D: That's what we're trying to produce isn't it – code! Of course it's best! (note: these presuppositions define the edge of the existing map)
A: Suppose we could generate the code...might that be more efficient? (note: this is the start of the new map)
D: I guess... maybe... I already use Generative AI to generate some code.
A: How do you do that?
D: What? I write prompts and ask the AI to generate code from them.
A: Would you like a way to generate better prompts?
D: Sure!
A: Suppose I told you that UML could help you to write better prompts, or even feed into an AI directly. Would that be interesting?
D: I suppose it might – but I would have to see it in action.
A: Would you be willing to do try it if we booked some training time to show you?
D: Sure – I guess...
In this case, It doesn't give us anything. It's best just to write the code is a presupposition that points to the edge of the developer's map. These statements are axiomatic within the boundaries of that map, so our next Question, Why is it best to just write the code? doesn't really have answer because it is a Question that is never supposed to be asked. We next introduce the possibility of code generation from UML using Generative AI, which is the new map. This implicitly answers the question Why is it best to just write the code? with the answer maybe it isn't, which begins to collapse the old map.
The key thing here, is that at every stage of the process, we must respect the other person's map. Otherwise, we just get into an argument. With a bit more work, we could probably convince the developer to make the paradigm shift from writing most code by hand to generating most code from higher level abstractions such as UML models. Developers tend to be very open to this sort of thing once they understand the benefits, but it can still be quite challenging in some cases depending on their level of emotional and ideological investment in the old map.
Another example where you might need to install a new map is to get stakeholders to accept a new system or a new way of doing things. This is often a management issue, but the Generative Analyst may find that old maps need to be replaced to get the information they need in stakeholder interviews. Often, you can just extend a map, but sometimes you really need a paradigm shift to get anywhere at all.