# topdown bottomup approach in complex systems

In continuation to question Topdown bottom up reductionism emergentism my second question is ; in complex system, like the economy, which are highly complex systems where it is often impossible to explain macro phenomena by a simple summation of micro processes or events. It’s hard to predict the results of actions in the system and sometimes impossible to find the causes of large abnormalities, or even to single out factors that influenced an event. So does top-down bottom-up approach does not work in complex systems ?

• The title of this question is very strange to me. Normally "top down" and "bottom up" are contrasted as two different approaches to understanding something. This question is the first time I've ever seen it phrased as if "top down bottom up" is one single approach.
– TKoL
Commented Mar 4 at 8:53
• Edison tried 1600 options before lucking into carbonized bamboo as a filament material Bottoms up. In contrast, Nuclear bomb #1 (trinity), worked impressively well, first try. Top-down, The top-down approach to light-bulb filaments wasn't panning out. Edison took a circuitous "shortcut". Commented Mar 4 at 9:56
• It is never possible to explain macro phenomena by a "simple summation" of micro processes or events, one has to account for their interactions. And it is hard to deduce non-trivial mathematical theorems generally, it does not mean that deductive approach does not work. Causes of large abnormalities and factors that influence an event may be too numerous to survey, without one or few dominant ones, but that is not the approach's fault. It is convenient for us when there is a single major cause we can focus on, but the world does not always oblige. Commented Mar 4 at 19:03

If you want to solve a set of differential equations do you need to understand the equations describing the local interactions or do you just need the boundary conditions (representing the global, outside conditions)? Obviously you need both. So it is rarely a "either/or" but often both: The rules that govern a system at the bottom and the outside constraints on the top, even in these extremely simple cases.

And then for more complex systems in order to understand them or have a chance to simulate them you need multiple levels in between. E.g. think of the task of understanding a complex biological system just from the basic equations of physics.

As for the economy you will have lots of things that are outside of what a narrow view on economics can predict. Global politics, new technical and scientific discoveries that significantly change some fundamental rules of the game...

The topic of Complexity targets not a precise domain (here you have an example of past answer of mine). Complexity refers mostly to a set of tools that serve to address systemic complexity (yes, it is circular). Consider here that complex means difficult to understand. If you try to address Complexity as a formal system, you will get lost in its contradictions.

The Complex Systems discipline is an attempt to differentiate "complex" systems from all the rest, that is, "simple" systems. However, the General Systems Theory born in fact to address the problem of complexity: to divide a complex problem in simple parts, so that solutions to parts can be found (a.k.a. Systemic Thinking in pop/mainstream jargon). If you don't understand a complex system, you can understand the simple components and even if you can't comprehend the totality (it is normal, we are limited beings), you can anyway manage it. The hardest part of the Systemic approach is the division of complex parts into simple parts: it requires a lot of experience and knowledge, it is not just turning the Systemic Thinking belly button switch on.

The top-down or bottom up approaches apply indistinctly to simple or complex systems (whatever that means). In both cases, the whole is divided in parts and it is not the whole who is addressed, but moreover the parts. But the core of the problem lies here: the parts MUST be simple.

• complex adaptive system has to have diverse connectedness interdependent adaptation Commented Mar 4 at 8:49