Post your week 8 discussion questions here as comments.
Week 8 Discussion Questions
5 Responses to Week 8 Discussion Questions
Leave a Reply
You must be logged in to post a comment.
Departments of Geography & Civil and Environ Engineering, UC Berkeley
(What) does a hierarchical patch dynamics paradigm have to contribute to an understanding of complex social systems? Can you think of examples or ways in which it might be applied or modified to be useful?
Back to basics: Why are complex systems usually hierarchically organized? What are the arguments put forth?
Wu and David posit that “In ecology, the hierarchies we construct inevitably result from the interactions between the inherent characteristics of the system under study and the observer who studies the system. While there is no absolute objectivity, how closely a constructed hierarchy corresponds to the structure of the real system significantly affects the usefulness and power of using a hierarchical approach” (9). This is quite a big challenge. Are the methods they cite to identify appropriate hierarchies sufficient? How does one negotiate this challenge of subjectiveness?
Both Lane and Wu & David discuss Simon’s concept of decomposition. How would the alternate perspectives of Anderson’s concept of macrolaws and Ginzburg’s “god in the details” stand in conversation with HPDP? What are the implications for treating systems as (nearly) decomposable?
Holland asserts (and Lane agrees) that “Persistent patterns arising from agent interaction are the key to emergence, which is the key to the formation of new levels of organisation” (117). Back to basics again: What is the connection between emergence and pattern formation?
In the Anderson theory (Lane, p14) is a simetrical broke of the stationary state similar to a bifurcation?
How can you built a model of a natural phenomenon when the sum of the parts (your units) are more than the simple sum of each single part?
Why in the theory of broken symmetry states (Lane, p14), the state of a very big system has less symmetrical laws than its constituents?
Which are the form of knowledge that doesn’t have a previous hierarchical structure? Are the discoveries in this category? Or are the discoveries parts of a mechanism of an emergent pattern?
Week 8: Hierarchy theory discussion questions for 3/12/13
1. “What is the degree to which system level properties simply emerge from local evolutionary forces and the degree to which those local processes are influenced and shaped by their effect on the persistence and continued functioning of ecosystems?“ Lane (2006) relates the connection of hierarchy and complexity theory using a history of prominent scientists explaining their ideas. How would Holland, Anderson, and Simon approach this question?
2. Holland defines a class of models–constrained generating procedures (cpgs)–in which simple rules generate both perpetual novelty and persistent patterns, or emergent phenomena. When cpgs produce “perpetual novelty” Holland says their state vectors tend to remain inside relatively small subsets of their history spaces. What does perpetual novelty really mean?
3. Explain the “probability-of-evolution” argument and why Holland uses it to suggest persistent patterns may contribute to emergence and generate new composite building blocks. Can we come up with another example (Holland uses the mammalian eye) of where emergence could contribute to a new level of hierarchy?
4. When identifying and developing models for spatially explicit patch hierarchy levels (Wu and David, 2002), how can we avoid the modifiable areal unit problem—when the same basic data yield different results when aggregated in different ways? When might it matter where we draw boundaries?
1. Is scale synonymous with a hierarchical level? Wu and David state that an important implication of vertical decomposition “is that the short-term dynamics of subsystems can be effectively and justifiable studied in isolation by ignoring the between-subsystem interactions that operate on significantly longer time scales.” Then they immediately follow with the caveat that “occasional exceptions to this general rule do exist as certain non-linear effects” that cross levels (or scales?). Are non-linear effects and cross-scale (=cross-level?) interactions essential properties of complex systems? If so, does that make hierarchy theory and this idea of short-term justifiable isolation moot — what do short-term, isolated studies tell us about complexity or emergence?
2. Is the above view of vertical decomposition reductionist? How does this differ from the contrasting constructivist point of view? How is SOC reductionist? Lane also discusses reduct. vs construct. on pages 92-94. Then he introduces that Holland’s idea of emergence is the “inverse of reductionism.” Who is on what side and should I care?
3. How loose is loosely-coupled? Are loose horizontal and vertical coupling valid assumptions? Wu and David state that “landscapes can be perceived as near-decomposable, nested spatial hierarchies in which hierarchical levels correspond to structural and functional units at distinct spatial and temporal scales.” Yet in this discussion, they emphasize the recognition of scale multiplicity as key to successful simplifications of complex systems. In their model of Pheonix land use, they don’t isolate scales, they use the model to “facilitate multiple scale modeling whereby local processes may be modified by top-down constraints or driving functions, and at the same time, bottom up propagation of info” hierarchically links interactions between scales. Are these contradictions of their earlier statement that short-term, isolated studies of subsystems are justified/valid? Or are these just examples of the importance of considering larger boundary conditions and averaging local conditions.
4. What is the evidence for the main tenets of the hierarchical patch dynamics paradigm? Do pattern and process only interact when operating in the same domain of scale in space and time? Do they really mean the same domain or just some overlapping domain? Aren’t there many examples of large scale processes overwhelming patterns and processes at a very small scale? Global CO2 levels on photosynthetic process, for example.
5. Are Wu and David’s scale breaks similar to discontinuities discussed in Allen et al (2005). Do the aggregated discontinuities that Allen et al studied (the body mass measurements, for instance) indicative of hierarchical levels? Can they indicate patch hierarchies?
6. Is Wu and David’s “triadic structure of hierarchy in research” valid: “When one studies a phenomenon at a particular hierarchical level (level x), the mechanistic understanding comes from the next lower level (level x-1), whereas the significance of that phenomenon is revealed at the next higher level (level x+1)?” Are LaTour’s observations or Lane’s idea of tangles more realistic? Is Wu and David’s structure just easier/cleaner to implement in research? In Lane’s conclusion (p118), he calls out triadic hierarchy in relation to “scaffolding structures.” Is this a similar concept?
7. In Wu and David, could analysis by a person produce the same, equally valid or better predictions of PHX’s future land use or veg response to flow regime than the hierarchical patch dynamics models? What are the benefits of using the programmed rules? Were any emergent patterns/properties discovered?
8. Lane suggests that from Holland’s perspective “modeling is the lynchpin of the scientific enterprise (pg 100).” He then follows by saying that modeling is a constructivist approach (is it, if it’s a set of simplified rules?) that allows us to identify emergent phenomena (is that the constructivist part?) that then requires proof “rather than merely simulated evidence.” Is this contradictory? What are we trying to prove anyway: how emergent phenomena arise? If we can define the rules and component interactions which give rise to emergent phenomena and simulate them, what more is there to prove? The causal relationships? Are there examples of studies which do this (Lane notes a lack of examples on page 104)? Or is the aim to “explain the origin of hierarchical levels of organization?” Chicken or egg?
9. How does Anderson’s “more is different” approach differ from Simon’s decomposability? It’s not sinking in, particle physics discussion not helping. Is this just about the hierarchy of the sciences? If so, is this silly (it’s just a social construct) or meaningful (relevant outside of academic distinctions)?
1. Do macro-laws and micro-laws operate within levels within a hierarchy, or do they also describe interactions between levels?
I think I asked about this before, but without the same terminology:
2. Lane describes “sandwiched emergence” as a process in which intermediate levels in a hierarchy are produced as the hierarchy is being established. Does such sandwiched emergence characterize tenuous interactions between levels, which become more established as the population of individual levels becomes more diverse? How does emergence function in relation to “richness”?
3. Lane also indicates that “tangled hierarchies” better characterize CES than ordered hierarchies; how does hierarchy theory then assist us in understanding CES, if we acknowledge that systems aren’t arranged strictly by level?
4. Can you further clarify commonly used modeling techniques in hierarchy theory (Wu and David, and any comparable models)?
5. How difficult is it to accurately model human-environmental interactions in models such as HPDM-PHX? How does this affect policy decisions?