What is a system?
- P.E. James: a system is "a whole (a person, a state, a culture, a business) which functions as a whole because of the interdependence of its parts". Geographers have used system ideas since the beginning, but had no technique for analysing complex systems until after WWII.
- Geography deals with multivariate, complex relations of living and non-living things in an ecosystem, so system analysis suits it well.
- Berry and Chorley proposed system analysis and general system theory as basic tools of geographic understanding.
- Chorley (1962): advantages of system analysis:
- Need to study systems rather than isolated phenomena
- Need to identify the basic principles governing systems
- Value of arguing from analogies
- Need for general principles covering various systems
General System Theory (GST) ⭐
- Developed by biologists in the 1920s, chiefly Ludwig von Bertalanffy: an organism can be understood only as a system of many associated parts. Later he saw the same idea applies to non-biological systems, so one analytical framework serves all sciences (General System Theory, 1968).
- A general system = a higher-order generalisation of many systems, a way of unifying the sciences that led to interdisciplinary research. GST is a theory of general models.
- Mesarovic: GST is concerned not merely with isomorphism and analogy but with a general theory from which the characteristics of systems can be deduced (deductive unification).
- It provides a unified hierarchical structure relating systems, stating when one approximates another, and even identifying useful system types not yet observed.
Cybernetics
- From the Greek kybernetes = helmsman: the study of regulating and self-regulating mechanisms in nature and technology (e.g. body temperature regulation). It follows common, mathematically describable laws. Regulation is precise in nature, defective in human societies.
- It emphasises interaction rather than a sharp cause–effect split. Causation can run both ways: an impulse returns to its origin after passing through other parts (feedback).
Open vs closed systems ⭐
- An open system interacts with surrounding systems, so it is hard to analyse. All real systems (e.g. landscapes) are open.
- For analysis a system must be treated as "closed": we consider a finite set of elements and relations, and assume excluded elements do not affect the system.
- A system is not the same as the model we make of it.
- Boundary problem: we must first set the system's boundaries. Harvey's firm example: analyse a firm as a closed system with fixed economic conditions. Extending the boundary to include changing social and political relations may change the result.
- Modelling a company: elements a₁ … aₙ (head office and branches) plus a₀ = the environment. Then study the connections between them.
- A system consists of: (1) a set of elements identified with variable attributes of objects; (2) a set of relationships between these attributes and the environment.
Merits of the abstract (system) approach
- Reduces complexity of a region to a simpler form for building models.
- Allows an abstract theory not tied to any one system.
- Gives information on possible structures, behaviours and states.
- Technical apparatus for interactions within complex structures.
- An abstract mathematical language (like geometry or probability) for empirical problems.
Structure of a system ⭐
Three components: (1) a set of elements, (2) a set of links, (3) links between the system and its environment.
Elements
- The basic aspects: structure, function, development.
- Structure = elements + connections.
- Function = flows (exchanges) along the connections.
- Development = change in structure and function over time.
- Mathematically an element is a primitive, undefined term (like a point in geometry).
- Scale-dependent: the international monetary system (countries) → an economy (firms) → an organisation (departments) → a department (people) → a person (a biological system). A car is an element of the traffic system and also a system itself.
- Blalock and Blalock (1959): systems A and B may interact as units, or their subsystems may interact at lower levels.
- Identification is easy for discrete units (farms) and hard for continuous ones (precipitation). An element should be construed as an attribute (variable) of an individual, not the individual itself.
Links (relationships) ⭐
Relation | Description | Example |
1. Series | Simplest: an irreversible link aᵢ → aⱼ, the classic cause and effect | Rice productivity in Punjab depends on irrigation. Saffron in Kashmir due to karewa soil |
2. Parallel | Two or more elements affect a third, or one affects two or more | Precipitation and temperature affect vegetation, and vegetation in turn affects rainfall and temperature |
3. Feedback | An element influences itself, directly (A→B→A) or indirectly via a chain | Leguminous crops enrich soil nitrogen, which benefits the crops |
4. Simple compound | Combination of the above | — |
5. Complex compound | Many combinations: a "wiring system" | — |
- Negative feedback: keeps a steady state through self-regulation (homeostatic / morphostatic). E.g. spatial competition progressively reduces excess profits until equilibrium.
- Positive feedback: the system changes its characteristics (morphogenetic), e.g. B’s effect on C causes further change in B via D.
Behaviour of a system
- Interrelationships of elements: flows, stimuli and responses, inputs and outputs. Internal behaviour = functional laws. External = transactions with the environment.
- A change in the environment affects one element, and the effect spreads to all connected elements: a simple stimulus–response (input–output) system without feedback, described by deterministic or probabilistic equations.
Geographical systems
- A system where one or more functionally important variables are spatial (location, distance, extent, density).
- Applications:
- Chorley: geomorphology as an open system (1962)
- Leopold and Langbein: entropy and steady state in fluvial systems
- Berry: "cities as systems within systems of cities" (organisation and information)
- Woldenberg and Berry: central-place and river patterns
- Curry: settlement location
- Haggett: locational analysis
- Static or adaptive systems are easy to build. Dynamic systems need time and space in one model, which is hard (see Lund's time–space model, Hägerstrand).
- Controlled systems: useful in planning when objectives and inputs are known. Usually only partially controlled: e.g. we can control fertiliser but not climate.
- Warning: the positive feedback of technology has driven exponential growth in population and industry. The long-term suppression of natural negative feedback may cause a crisis of pollution, hunger and resource shortage.
- Practical research need not use the full mathematics. Iron-ore example: elements = producing and consuming centres; links = trade lines; function = tonnage moved; development = maps over time.
- The system approach was technically demanding, so it attracted fewer researchers.
- Criticism: tied to positivism, ignoring normative values (aesthetics, beliefs, attitudes, hopes, fears), so it gives no real picture of a region's personality.
Stages of development of a science
Stage | In geography |
1. Descriptive (description and mapping) | Antiquity to the mid-18th century |
2. Analytical (explanation, laws) | The Humboldt period; analysis of spatial distribution began |
3. Predictive (models predict) | Partly reached with geomorphology and climatology in the late 19th century. In human geography after WWII (predictive locational theories, controlled-system models). Geography is now in the predictive stage |
Haggett's internal structure of geography ⭐ (Geography: A Modern Synthesis)
Haggett divided geography by mode of analysis (rather than physical vs human and so on):
Analysis | Concern | Theoretical side | Applied side |
1. Spatial analysis | Variation in the location and distribution of phenomena (rainfall, wheat yield) and its controlling factors | Spatial interaction theory, diffusion theory | Watershed development, urban problems |
2. Ecological analysis | Human–environment connections in a closed or partly closed ecosystem (intra-regional). Linked to homogeneous regions | Environmental structure, ecosystems | Natural resource geography, hazard appraisal |
3. Regional complex analysis | Combines spatial and ecological analysis. Regions identified by areal differentiation, then inter-regional flows and feedbacks studied. Based on functional regions (centre + tributary area, boundary where the centre's influence equals another's, nested hierarchically) | Regional growth theory, inter-regional flow theory | Regional forecasting, regional planning |
- Supporting areas: methodology, philosophy, history of geography; techniques (cartography, quantitative methods); systematic (physical: geomorphology, climatology; human: economic, urban) and regional (tropical, arid lands, Latin America, South Asia) fields.
- Modern geography is a science of synthesis, yet regional synthesis has made little progress, because physical and human components follow different laws. New philosophies (positivism, idealism, realism, dialectical materialism) offer partial answers.