Dynamics Systems Theory (DST):
- Emphasis on time and how systems change over time
- Giunti : Real dynamical system - real system changes over time, Mathematical dynamic system - system of equations changing over time
- Consists of state space, trajectories and attractors
	- STATE SPACE : Hypothetical space with 1 dim per variable describing eqn of system
	- TRAJECTORY : Joining the "dots" of individual states in the state space over time 
	               forms a line called trajectory 
	- ATTRACTOR : Plotting several years' worth of data might form clusters of 
		      trajectories at particular intervals of time - called attractors. 

		      Looks like that region of the state space is attracting a trajectory towards
 		      itself - variety of starting points converging to the same final state. Indicates repetitive pattern 
		      of negative feedback loop (eg. in people with depression - negative feedback loop of self-worth, 
		      shows massive canyon in neural dynamics - steady state (why steady?, habitual thought patterns) Is settling at the point considered to 
	              attract other future states to come back to that state or are we claiming that that is a stable state because 		      
		      states keep coming back to that point)
	- FEEDBACK LOOPS : Parts of a complex systems tend to couple together - like 2 loops of Lorentz attractors
	              

AEON - STICKY THOUGHTS, STICKY DYNAMICS	

		      "A point or area in the state space that the system keeps gravitating towards"

	- PHASE PORTRAIT : Corresponds to a complete picture of the different tendencies a
 			   system might have 

- Nature hypothesis v/s knowledge hypothesis :
	- NATURE hypothesis - Claim that cognitive systems are dynamic rather than
 			      computational 
	- KNOWLEDGE hypothesis - Cog Sci should proceed dynamically (using tools of DST)

Van Gelder : 
	- Undermine 'only game in town' narrative of GOFAI (attributed to Fodor)
	- Cognitive Systems as dynamic systems rather than static Turing Machines

Examples of Dynamic Systems : 
	- Population Dynamics
	- 


Criticises atemporal functionalism and GOFAI
- GOFAI and Comp func tend to treat cognitive processes as sequences of discrete, step-by-step operations -> focus on ORDER OF OPERATION 
- Both hypotheses of the dynamical approach to cognition accept the time-dependent nature of cognitive processes. Previous models of cognition individuated mental states based on their causal roles. Functionalism is often thought of as atemporal - ignoring the time-dependent nature of cognitive processes. Even critics of the dynamical approach - Eliasmith (1996) - agree that functionalism could do with an added temporal aspect. 
- [Van Gelder & Port, 1995] Computational models specify a discrete sequence of static internal states in arbitrary 'step time' (ersatz time). Cognitive processes unfold continuously in real time. 
- GOFAI relies on internal representations - cognition relies on processing these representations to generate thoughts, decisions or behaviours. DST suggests that representations may not be necessary for intelligence. Instead of manipulating internal symbols, DIST sees cognition as a direct, real-time interaction between an agent and its env. 




QUESTIONS : 
> Neural systems tend to / are likely to revisit those states where attractors are created - Attributing intentionality to neural states?? 
- You're not assigning intentionality to the neural states, you're assigning it to yourself. The neural state of depression (that is the attractor) doesn't arrive at itself, you are likely to approach that state under certain conditions if that state has been reinforced as an attractor in your mental model