Jun 2007

How to Approach Hard vs. Soft ProblemsTwo problem solving approaches: Holism vs. Reductionism

Let’s preface this discussion by asking a fundamental question

What is Intelligence? What is it used for?

The purpose of intelligence is for prediction

● Intelligence is for prediction

● Prediction is a low level operation in the brain

● Prediction not logic is most important

Many complex systems including entrepreneurial ventures and creating hit entertainment products require prediction as a

fundamental skill set to achieve success

Throughout history two fundamental approaches to understand science and the world around us have been used: Reductionism and Holism

Reductionism Holism

● Parts, Division ● Context, Whole, Environment

● Math, Physics, Computer Science ● Biology, Ecology, Philosophy

● Programmers, Surgeons, Engineers ● Nurses, Authors, Philosophers

● Proof, Precise Measurement, Prediction ● Categories, Description, Speculation

Today we live in a world ruled by Reductionism and Reductionist scientific approaches

Reductionism vs. Holism

Reductionism focuses on Component Dominated Complexity

Reductionist Approach to Complex Systems

System

Component 2 Component 3

Sub-Component

Sub-Component

Sub-Component

Solution for System Complexity

● Manage complexity through division● DIvide the system into components● Create simple interfaces between components

Component 1

Holism on the other hand, focuses on Interaction Dominated Complexity

Holistic Approach to Complex Systems

Examples

● Neurons in the brain

● People in society

● Concepts, abstractions, ideas

Chaotic Systems

Chaotic Systems and Reductionism

● Stateful components

● Non-linear components

● Interaction dominated complexity

● Chaotic systems are common in life

● Non-divisible complexity

● Can’t use reductionist science for prediction

Chaotic Systems Characteristics Key Insights

Ambiguity in Systems

Overview

➢ Incomplete information

➢ Self reference, loops

➢ Chicken and the egg problem

➢ Incorrect information○ Lies, misunderstandings○ Multiple points of view, opinions○ Persuasion

Irreducible Complexity in Systems

Overview

➢ Emergent properties

➢ Everything matters○ Internally: Curse of Dimensionality○ Externally: Can’t separate “system” from environment

PROCESS

Complex Systems that defy Reductionism

1. Chaotic

2. Contain Ambiguity

3. Irreducible Complexity

4. Require a Holistic Stance

We have described four kinds of complex systems that defy Reductionism and are unpredictable relative to reductionist approaches

Soft sciences are more difficult because soft science tends to deal with more complex systems than hard science does

Overview

➢ Soft science cannot make as good prediction as hard sciences because they have to deal with life

➢ Life is bizarre

➢ Reductionist (Hard) science cannot deal with bizarre systems

➢ Reductionist success comes from limiting their problem down to non-bizarre systems

We can express various classes of problems based on the amount of complexity of the system and the range of prediction possible

Complexity and Prediction

Examples of Bizarre Systems

➢ Entrepreneurial ventures / Venture capital

➢ Language translation

➢ Weather

➢ Stock markets

➢ Human interest / intent / recommendations

➢ Internet search

➢ Hit mobile game design & development➢ Etc., etc.

Today Reductionist science has solved a major class of problems in the Complexity/Prediction graph

Complexity and Prediction

Key Takeaway: Different classes of problems require different approaches to solve!

Complexity vs Prediction Problem Classes