From Diagnosis to Causal History
A Temporal Causal-Reconstruction Framework for Distinguishing Predisposition, Initiation, Transmission, Amplification, Terminal Mortality, and Decomposition in Tree Decline
John Swygert
September 23, 2026
Hypothesis, methods, and research-framework paper
Abstract
Tree decline is often diagnosed from the biological agents, injuries, physiological abnormalities, or environmental conditions detectable when symptoms become conspicuous. Such diagnoses can correctly identify an important causal agent without necessarily reconstructing the sequence by which the tree entered decline. A fungus capable of causing lethal disease may have initiated a particular decline, may have entered after insect injury, may have become damaging only after environmental or physiological stress, or may have proliferated after the tree had crossed an irreversible threshold. Likewise, an insect may initiate injury, vector a pathogen, amplify an existing disease, exploit a weakened host, or kill directly. Similar ambiguity applies to bacteria, viruses, other microorganisms, parasites, microbiome changes, drought, heat, flooding, nutrient disruption, pollution, root injury, mechanical damage, host genetics, symbiotic disruption, hydraulic dysfunction, carbon limitation, and interacting processes.
This paper develops a Temporal Causal-Reconstruction Framework (TCRF) for distinguishing these possibilities. The framework separates four questions that are frequently conflated: what is present; what can cause disease or injury; what occurred first; and what materially changed the trajectory toward decline or death. It further separates causal description along four dimensions: temporal position, biological mechanism, causal contribution, and organizational scale. Rather than requiring one universal initiating agent, the framework treats insect-first, fungus-first, bacterial-first, viral-first, environmental-stress-first, microbiome-first, injury-first, soil/root-first, physiological-first, simultaneous-initiation, and feedback-driven models as competing hypotheses.
The central methodological proposal is that terminal diagnosis should not substitute for causal history. Candidate histories should be specified before confirmatory analysis where possible and tested using longitudinal observation, factorial manipulation, exclusion and rescue experiments, molecular detection, tree physiology, microbiome profiling, dendrochronology, historical reconstruction, remote sensing, and causal modeling. Temporal priority alone is insufficient for causation, while demonstrated pathogenic capacity alone is insufficient to establish initiation of a particular decline.
The framework is compatible with established forest pathology and does not challenge demonstrated disease mechanisms merely because an earlier event may exist. Instead, it asks whether the recognized agent explains initiation, transmission, amplification, terminal damage, or some combination of these roles.
A final section translates the framework into The Swygert Theory Of Everything AO (TSTOEAO) as a prospective test rather than an imposed biological explanation. Encoded Equilibrium, boundary conditions, channel-selective routing, structured response, and recursive boundary construction may provide testable descriptions of state-dependent susceptibility and path dependence only when biological variables are independently measured, predictions are registered before outcomes, and conventional explanations remain explicit comparators.
Keywords: tree decline; forest mortality; causal inference; forest pathology; forest entomology; microbiome; fungi; insects; bacteria; viruses; drought; hydraulic failure; causal sequence; TSTOEAO; Encoded Equilibrium.
1. Introduction: A Diagnosis Is Not Necessarily a History
A declining tree is a historical object.
By the time crown thinning, cankers, fungal structures, galleries, vascular discoloration, root deterioration, foliage loss, or structural decay become conspicuous, the tree may contain the accumulated consequences of months, years, or decades of interacting events.
An investigator arriving at that stage sees an endpoint.
The endpoint matters. It may contain a pathogen whose causal capacity is experimentally established. It may reveal an insect capable of killing the host directly. It may show drought injury, root disease, nutrient deficiency, mechanical damage, or another well-characterized process. Nothing in the present framework requires those observations to be discounted.
The methodological problem arises when evidence about the endpoint is treated as though it necessarily reconstructs the beginning.
The distinction can be stated simply:
What is causing damage now is not necessarily identical to what first shifted the system toward decline.
The reverse error is equally important. An event that occurred first is not necessarily the principal cause of later mortality. Temporal priority does not by itself establish causal importance.
A tree may therefore have more than one scientifically valid causal description.
A pathogen may cause the disease that ultimately kills the tree while an earlier drought episode materially increased susceptibility. An insect may introduce a fungus without first producing a prolonged period of physiological weakening. Root disease may precede insect colonization. An insect may attack first but contribute little to eventual mortality. A microbiome shift may follow physiological deterioration rather than cause it. Environmental stress and infection may interact so early that searching for one unique initiator becomes biologically misleading.
The purpose of this paper is therefore not to replace conventional diagnosis.
It is to add another question:
What causal history produced the diagnosed state?
This question extends earlier work in which insect initiation was proposed as a testable pathway in some forest declines. That earlier hypothesis distinguished physical injury, pathogen delivery, and physiological or microbial disruption, while acknowledging fungal initiation, environmental stress, direct insect mortality, and combined pathways as alternatives. The subsequent global framework broadened the question further, treating insect-first, pathogen-first, stress-first, injury-first, simultaneous, and feedback-driven pathways as competing hypotheses. The open causal framework then removed any preferred initiating category and explicitly included fungi, insects, viruses, bacteria, other microorganisms, parasites, environmental stress, physical injury, soil and nutrient conditions, host genetics, physiological state, and symbiotic relationships.
The present paper takes the next methodological step.
It asks how competing causal histories can actually be distinguished.
2. The Central Causal Distinction
Four propositions should be separated.
Proposition A: Presence
Agent or condition X is detectable in association with the declining tree.
Proposition B: Pathogenic or damaging capacity
Under appropriate conditions, X can produce injury, disease, dysfunction, or mortality.
Proposition C: Temporal priority
X occurred before another candidate event Y.
Proposition D: Causal initiation
Occurrence of X materially changed the probability or trajectory of subsequent decline.
These propositions are related but not interchangeable.
Detection establishes presence.
Experimental pathology may establish pathogenic capacity.
Longitudinal observation may establish temporal order.
Causal initiation requires stronger evidence that the trajectory would probably have differed in the absence of the candidate initiating event.
Accordingly:
[ X \text{ can cause disease} ]
does not logically entail
[ X \text{ initiated this decline}. ]
Likewise:
[ X \text{ occurred first} ]
does not entail
[ X \text{ caused the later outcome}. ]
The strongest question is counterfactual:
[ P(D \mid do(X=x_1),Z) \neq P(D \mid do(X=x_0),Z), ]
where D represents a prespecified decline outcome, X the candidate causal factor, and Z relevant registered conditions or covariates.
For historical events where direct intervention is impossible, this relation cannot simply be assumed. It must be approximated through natural experiments, longitudinal evidence, mechanistic reconstruction, causal models, historical records, matched comparisons, sensitivity analysis, or converging independent evidence.
The framework therefore distinguishes causal capability from case-specific causal history.
3. From a Single Cause to a Causal Trajectory
Tree decline can be represented as a sequence of changing biological states:
[ S_0 \rightarrow S_1 \rightarrow S_2 \rightarrow \cdots \rightarrow S_n, ]
where S_0 is the registered or reconstructed initial state and S_n is the observed terminal state.
The transition between states may depend on multiple processes:
[ S_{t+1}=F(S_t,B_t,E_t,P_t,M_t,I_t,G,H_t,\varepsilon_t), ]
where, depending on the study:
B_t represents biological agents;
E_t represents environmental conditions;
P_t represents host physiological state;
M_t represents microbiome or symbiotic state;
I_t represents injury or disturbance;
G represents host genetic characteristics;
H_t represents historical or accumulated state;
\varepsilon_t represents bounded unexplained variation.
This equation is a conceptual state-transition representation, not a claim that these variables are presently measurable as one universal function.
Different tree declines may follow different trajectories.
Insect-first
[ I \rightarrow Injury \rightarrow Infection \rightarrow Decline ]
Pathogen-first
[ Pathogen \rightarrow Tissue\ Damage \rightarrow Insect\ Exploitation \rightarrow Decline ]
Environmental-stress-first
[ Drought/Heat \rightarrow Physiological\ Stress \rightarrow Susceptibility \rightarrow Biotic\ Damage ]
Root-first
[ Root\ Injury/Disease \rightarrow Reduced\ Uptake \rightarrow Crown\ Stress \rightarrow Secondary\ Attack ]
Microbiome-mediated
[ Community/Functional\ Change \rightarrow Host\ State\ Change \rightarrow Altered\ Susceptibility ]
Vector-mediated
[ Vector \rightarrow Pathogen\ Delivery \rightarrow Disease ]
without requiring a prolonged weakening phase.
Simultaneous interaction
[ A+B \rightarrow Decline ]
where neither factor independently reproduces the observed outcome.
Reciprocal amplification
[ A \rightarrow B \rightarrow A^{+} \rightarrow B^{+} \rightarrow Collapse. ]
These are candidate structures, not universal laws.
4. The Expanded Causal Field
A causal-reconstruction framework should not begin by deciding which class of factor deserves priority.
Candidate initiators and modifiers include, where biologically relevant:
fungi; oomycetes; insects and other arthropods; bacteria; viruses; viroids; phytoplasmas; nematodes and other parasites; endophytes capable of state-dependent behavior; altered microbial communities; disruption of mycorrhizal relationships; drought; heat; cold; flooding; altered vapor-pressure deficit; fire; wind; lightning; freeze–thaw injury; salinity; atmospheric pollution; ozone; nitrogen deposition; chemical toxicity; herbicide or pesticide exposure; nutrient deficiency or imbalance; soil acidification; compaction; altered drainage; erosion; root severance; construction disturbance; browsing; mechanical injury; competition; stand density; age; tree size; phenological state; host genotype; provenance; local adaptation; hydraulic vulnerability; carbon allocation; reproductive load; and interactions among these variables.
This list remains deliberately incomplete.
An open framework must retain an unmodeled-cause category.
Otherwise every unexplained outcome risks being forced retrospectively into the existing causal vocabulary.
5. Four Axes of Causal Role
Terms such as initiator, vector, facilitator, accelerant, opportunist, terminal pathogen, and decomposer are useful, but a single label can conceal several different causal properties.
The present framework therefore separates four axes.
5.1 Temporal position
A factor may function as:
Predisposer — alters susceptibility before an identifiable decline episode.
Initiator — produces the first demonstrated material transition toward decline.
Propagator — carries the process between tissues, trees, or locations.
Amplifier — increases severity or transition rate after decline has begun.
Terminal mechanism — produces or participates in the transition across an irreversible mortality threshold.
Postmortem participant — acts predominantly after death or irreversible loss of function.
These positions may overlap.
5.2 Mechanistic role
A factor may operate through:
physical injury;
vectoring;
infection;
toxicity;
vascular obstruction;
hydraulic disruption;
carbon depletion or allocation change;
defense modification;
root impairment;
nutrient disruption;
symbiosis alteration;
resource competition;
hormonal or signaling disruption;
microbiome modification;
tissue consumption;
structural weakening;
or decomposition.
5.3 Causal contribution
Evidence may support classification as:
necessary;
sufficient under specified conditions;
contributory;
modifying;
correlated but causally unresolved;
or unknown.
These classifications must remain condition-specific. A factor can be necessary in one disease system but not in another, or necessary only under a particular host genotype or environmental regime.
5.4 Organizational scale
The same factor may have different causal significance at different scales:
cell;
tissue;
organ;
whole tree;
stand;
landscape;
regional epidemic;
or evolutionary population.
This distinction is essential because the cause of an individual tree's death is not necessarily identical to the cause of an epidemic.
6. Predisposition Is Not Initiation
The term predisposition requires particular care.
Suppose a drought reduces stored carbon, hydraulic safety margin, fine-root function, or defense capacity. Months later a pathogen infects the tree and produces lethal disease.
Was drought the cause?
Was the pathogen the cause?
The answer depends on the causal question.
If the pathogen is required for the disease, calling drought the pathogen's replacement would be incorrect.
If the probability or severity of infection would have been substantially lower without drought, calling drought irrelevant would also be incorrect.
The framework therefore allows layered causation.
A predisposer changes the probability landscape within which a later event acts.
An initiator marks the first material transition in the defined decline episode.
A terminal mechanism produces the final irreversible transition.
These can be different events.
7. Physiological State Must Be Treated as More Than “Stress”
The word stress is too broad to carry the full physiological problem.
Modern tree-mortality research identifies water balance, hydraulic conductivity, xylem embolism, carbon supply and demand, root-to-soil conductance, tissue dehydration, defense capacity, and their interactions as important mechanisms in drought-associated mortality.
Accordingly, causal reconstruction should measure physiological states wherever feasible rather than assigning a generic “environmental stress” category.
Candidate measurements include:
predawn and midday water potential;
sap flow;
hydraulic conductivity;
percentage loss of conductivity;
xylem vulnerability;
nonstructural carbohydrate concentrations;
photosynthetic rate;
stomatal conductance;
cambial activity;
fine-root production and mortality;
root-to-soil conductance;
defense compounds;
resin or latex production where relevant;
nutrient status;
crown temperature;
and recovery after disturbance.
A physiological transition may itself be the mechanism linking environmental exposure to later biological attack.
8. Microbiomes: Composition Is Not Function
Microbiome evidence presents a special risk of overinterpretation.
A tree contains multiple microbial communities associated with leaves, bark, phloem, xylem, roots, rhizosphere, surrounding soil, and other compartments.
A detected change in one compartment does not establish equivalent change in another.
Likewise:
[ \Delta Community\ Composition ]
does not automatically imply
[ \Delta Beneficial\ Function. ]
Nor does association establish causal direction.
A microbiome change could be:
an initiator;
a mediator;
a consequence of physiological decline;
a response to insect injury;
a response to pathogen invasion;
a compensatory response;
an opportunistic transition;
or an incidental correlate.
A strong microbiome claim therefore requires functional evidence whenever possible.
Manipulation and rescue are particularly valuable.
If disruption of community M increases decline and restoration of relevant microbial function reduces decline under controlled conditions, a causal interpretation becomes substantially stronger than one based solely on sequencing differences.
9. Latent, Endophytic, Opportunistic, and State-Dependent Organisms
The framework must also allow organisms to change ecological role.
The categories beneficial, commensal, endophytic, opportunistic, pathogenic, and saprotrophic should not always be treated as immutable properties of a species.
Host state, environment, microbial interactions, tissue condition, genotype, and resource availability may alter biological behavior.
This produces an important causal possibility:
[ Organism\ Present_{healthy} \rightarrow Host/Environment\ Transition \rightarrow Organism\ Pathogenic_{decline}. ]
In such a case, detection of the organism before decline would not establish that it initiated decline.
The relevant transition may instead be the change that altered the organism–host relationship.
10. Individual Mortality and Epidemic Causation Are Different Questions
Consider an introduced pathogen capable of killing susceptible trees.
For an individual tree, causal reconstruction may ask:
What sequence led this tree to infection and death?
At population scale:
What permitted the pathogen to spread through the population?
At landscape scale:
What environmental, host-density, vector, climate, disturbance, or connectivity conditions permitted epidemic expansion?
At evolutionary scale:
Why was the host population susceptible?
All four explanations may be correct.
They are not interchangeable.
A factor unnecessary for death after infection could nevertheless be critical for epidemic spread.
Conversely, a factor important in one individual's deterioration may be irrelevant to the regional epidemic.
The unit of causal inference must therefore be declared before interpretation.
11. American Chestnut as a Model of the Distinction
The American chestnut illustrates why causal levels must be separated.
The established role of Cryphonectria parasitica in chestnut blight should remain intact. The fungus is capable of producing destructive bark cankers, girdling susceptible stems, and driving severe disease.
A historical reconstruction can nevertheless ask additional questions.
What was the physiological and ecological condition of chestnut populations when the epidemic entered different landscapes?
Did disturbance history alter susceptibility?
Did preexisting root disease affect some populations?
Did insect injury alter infection probability in some trees?
Did soil or microbial conditions affect disease progression?
Did host genotype dominate these effects?
Were surviving mature trees disproportionately associated with particular environments?
These questions do not become evidence merely because they are plausible.
The earlier Forest Before the Blight hypothesis explicitly left microbiome preconditioning, insect initiation, survivor distribution, and microbiome-mediated protection unproven and proposed historical reconstruction and factorial testing.
The present framework places those propositions into competing causal models rather than combining them into one preferred narrative.
For example:
[ H_1: Fungus \rightarrow Disease \rightarrow Mortality ]
[ H_2: Root\ Disease \rightarrow Reduced\ Host\ Function \rightarrow Fungus \rightarrow Mortality ]
[ H_3: Insect\ Injury \rightarrow Infection\ Court \rightarrow Fungus \rightarrow Mortality ]
[ H_4: Disturbance \rightarrow Soil/Microbial\ Change \rightarrow Altered\ Host\ State \rightarrow Fungus ]
[ H_5: Host\ Genotype \rightarrow Susceptibility \rightarrow Fungus \rightarrow Mortality ]
[ H_6: Multiple\ Preconditions \rightarrow Fungus \rightarrow Mortality. ]
The historical evidence should determine which models remain viable.
Failure to support H_2-H_6 would not weaken the demonstrated pathogenicity of C. parasitica.
Conversely, support for a predisposing pathway would not require demoting the fungus from its demonstrated role in blight.
12. The Hidden-State Problem
Causal reconstruction faces a difficult asymmetry.
Early events are often easier to miss than late events.
An insect may leave.
A wound may heal over.
A transient viral infection may no longer be detectable.
A microbial community may reorganize.
Roots may decompose.
Drought may end before symptoms appear.
A physiological threshold may have been crossed months earlier.
Therefore:
[ Not\ observed \neq Demonstrated\ absent. ]
But the opposite safeguard is equally necessary:
[ Not\ observed \neq License\ to\ assume. ]
A proposed hidden event must remain falsifiable.
The statement “an undetected insect probably initiated the decline” cannot be protected indefinitely by arguing that the insect disappeared before observation.
Instead, hidden-event hypotheses should generate residual predictions.
For example, prior boring may leave anatomical traces.
Historical drought may be recoverable from meteorological records or growth patterns.
Past defoliation may leave growth signatures.
Previous vascular injury may alter wood anatomy.
Pathogen ancestry may be reconstructed genetically in some systems.
Historical imagery may reveal crown decline.
Archived herbarium, wood, soil, or museum material may preserve molecular evidence.
The invisible past becomes scientifically useful only when it leaves independently testable consequences.
13. Temporal Evidence
A causal-reconstruction program should assign evidentiary strength to different forms of temporal evidence.
Potential sources include:
prospective repeated sampling;
permanent forest plots;
tree-ring records;
wound and callus chronology;
gallery age;
lesion development;
canker growth;
root necrosis progression;
annual aerial imagery;
satellite remote sensing;
LiDAR;
thermal imaging;
historical photographs;
forest inventories;
weather records;
land-use records;
herbarium material;
archived wood;
museum insect collections;
stored soil;
environmental DNA;
pathogen population genetics;
host transcriptomic signatures;
and repeated microbiome samples.
Prospective evidence should generally receive greater weight for sequence reconstruction than retrospective inference from a terminal specimen, all else equal.
14. Mechanistic Evidence
Chronology alone cannot establish causation.
If insects consistently appear six months before a fungus, at least four explanations remain possible:
insect injury facilitates fungal establishment;
an unmeasured prior condition independently promotes both insects and fungi;
early fungal or physiological change attracts insects before the fungus becomes detectable;
the insect and fungus are temporally associated but the insect does not materially alter disease progression.
Mechanistic experiments are therefore required.
Useful designs include:
Exclusion experiments — prevent the proposed initiator while leaving other conditions as similar as possible.
Addition experiments — introduce the proposed factor under controlled conditions.
Rescue experiments — restore the proposed damaged function.
Crossed-factor experiments — manipulate multiple candidate causes independently.
Sequence experiments — change the order of exposure.
Dose experiments — determine whether intensity changes transition probability.
Timing experiments — vary phenological or physiological state at exposure.
Genotype experiments — test host-state dependence.
Microbiome-transfer experiments — where scientifically and biosafety appropriate, test whether defined microbial states alter outcome.
15. Sequence Must Become an Experimental Variable
One of the most important tests follows directly from the causal-history hypothesis.
Suppose two factors are implicated:
[ A = insect\ injury ]
and
[ B = fungal\ exposure. ]
Most factorial designs would test:
control;
A;
B;
A+B.
Temporal reconstruction requires more.
It should compare:
[ A \rightarrow B ]
against
[ B \rightarrow A, ]
and where appropriate:
[ A+B\ simultaneously. ]
The same principle applies to:
drought → insect versus insect → drought;
drought → pathogen versus pathogen → drought;
microbiome disruption → pathogen versus pathogen → microbiome disruption;
root injury → pathogen versus pathogen → root injury.
If sequence materially changes outcome while cumulative exposures remain comparable, temporal order itself becomes biologically informative.
16. A Prospective Forest Observatory
The strongest test of the framework would begin before visible decline.
A network of apparently healthy trees should be enrolled and repeatedly sampled across species, sites, ages, genotypes where known, soil conditions, disturbance histories, and environmental gradients.
Measurements should be scheduled before investigators know which trees will decline.
A core program could include:
insect trapping and direct inspection;
fungal detection;
bacterial profiling;
viral/metagenomic screening where feasible;
root and rhizosphere sampling;
mycorrhizal characterization;
soil chemistry;
soil moisture;
weather;
sap flow;
water potential;
hydraulic measurements;
nonstructural carbohydrates;
defense chemistry;
growth;
phenology;
crown condition;
thermal imagery;
and high-resolution photographs.
When a tree later declines, the study would possess something terminal diagnosis usually lacks:
a biological record of the tree before it became visibly sick.
Cases could then be matched against trees that remained healthy.
The central analysis would ask which measurable transitions reliably preceded decline, which followed it, and which altered subsequent transition probability.
17. Molecular Reconstruction
Molecular technologies expand the observable causal history but do not automatically solve it.
Potential tools include:
amplicon sequencing;
shotgun metagenomics;
metatranscriptomics;
host transcriptomics;
proteomics;
metabolomics;
pathogen genotyping;
viral discovery;
environmental DNA;
stable-isotope approaches;
and spatially resolved molecular sampling.
These methods can identify previously invisible biological states.
However, detection must not be confused with causal interpretation.
Finding viral nucleic acid does not establish viral initiation.
Finding fungal DNA does not establish active disease.
Finding a bacterial taxon enriched in declining trees does not establish bacterial causation.
Finding a microbiome difference does not establish functional disruption.
Molecular evidence should therefore be integrated with chronology, abundance, activity, host response, tissue localization, intervention, and outcome.
18. Counterfactual Causal Questions
For each candidate factor X, investigators should ask:
If X had not occurred, would the subsequent trajectory probably have remained the same?
This can be decomposed.
Initiation effect
Would decline have begun during the same interval without X?
Acceleration effect
Would decline have progressed at the same rate without X?
Severity effect
Would the maximum damage have been comparable without X?
Mortality effect
Would the tree have crossed the irreversible mortality threshold without X?
Transmission effect
Would the damaging agent have reached this host without X?
Population effect
Would the outbreak have spread similarly without X?
These are different estimands.
A factor can have a large transmission effect and a small post-infection severity effect.
Another can have no transmission role but dominate mortality once infection occurs.
19. Causal Graphs Before Narrative
Candidate causal models should be represented explicitly before they are converted into prose.
A simplified graph might contain:
[ Climate \rightarrow Host\ Physiology ]
[ Soil \rightarrow Root\ Function ]
[ Host\ Physiology \rightarrow Insect\ Susceptibility ]
[ Host\ Physiology \rightarrow Pathogen\ Susceptibility ]
[ Insect \rightarrow Injury ]
[ Insect \rightarrow Pathogen\ Delivery ]
[ Injury \rightarrow Pathogen\ Establishment ]
[ Pathogen \rightarrow Vascular\ Damage ]
[ Vascular\ Damage \rightarrow Physiological\ Stress ]
[ Physiological\ Stress \rightarrow Insect\ Susceptibility ]
[ Vascular\ Damage \rightarrow Mortality. ]
The feedback edge from physiological deterioration back to susceptibility makes the system path-dependent.
Competing graphs should be compared rather than allowing one narrative to absorb every result after observation.
Where cycles make a standard directed acyclic graph inappropriate, dynamic causal models, cross-lagged structures, state-space models, multistate survival models, or other time-indexed approaches should be considered.
20. The Irreversibility Threshold
A major unresolved problem in tree mortality is determining when decline becomes irreversible.
A tree may be damaged without being committed to death.
Therefore the framework distinguishes:
[ Declining ]
from
[ Irreversibly\ dying. ]
This distinction matters for causal attribution.
An organism appearing after irreversible failure may contribute to tissue destruction without having contributed to the transition into mortality.
The ideal study should therefore attempt to estimate a transition:
[ S_{recoverable} \rightarrow S_{irreversible}. ]
Candidate indicators may include hydraulic thresholds, cambial death, loss of viable meristems, unrecoverable root failure, or other species-specific physiological markers.
The appropriate threshold must be established biologically rather than assumed from the present framework.
21. Feedback Loops
Many declines may not possess a single linear chain.
Consider:
[ Drought \rightarrow Reduced\ Carbon\ Gain \rightarrow Reduced\ Defense \rightarrow Insect\ Attack \rightarrow Vascular\ Damage \rightarrow Reduced\ Water\ Transport \rightarrow Further\ Carbon\ Loss. ]
Or:
[ Root\ Disease \rightarrow Water\ Stress \rightarrow Crown\ Decline \rightarrow Insect\ Colonization \rightarrow Additional\ Vascular\ Damage \rightarrow Root\ Carbon\ Limitation. ]
Once such loops begin, asking which factor is “the cause” without specifying a time window may become misleading.
The scientifically useful questions become:
What entered the loop first?
Which edge is strongest?
Which edge is necessary for continued amplification?
Which intervention breaks the loop?
Which process pushes the tree across the irreversible threshold?
This converts vague multifactorial explanation into testable causal structure.
22. Distinguishing Opportunism From Terminal Causation
The term secondary can accidentally imply unimportance.
That should be avoided.
An organism may arrive late and still kill the tree.
A late-stage pathogen can therefore be temporally secondary but mechanistically lethal.
Conversely, an early factor may be temporally primary but weak in direct destructive capacity.
Accordingly, the framework rejects the equation:
[ Secondary = Unimportant. ]
Instead:
[ Temporal\ Position \neq Causal\ Magnitude. ]
The same principle applies to decomposers. Detection in dead tissue may identify organisms that consume the aftermath rather than organisms responsible for death.
The distinction requires tissue viability, colonization timing, pathology, host response, and where possible experimental inoculation or exclusion evidence.
23. Falsification Matrix
The framework becomes useful only if candidate sequences can lose support.
Insect-first is weakened when:
repeated prospective sampling shows fungal/pathogen activity preceding insect injury;
insect exclusion does not alter infection, progression, or mortality within prespecified equivalence margins;
insect injury fails to reproduce predicted physiological or susceptibility changes;
or pathogen exposure alone reproduces the complete trajectory.
Fungus-first is weakened when:
prospective sampling consistently shows damaging insect activity before fungal establishment;
fungal colonization requires experimentally demonstrated prior injury under the tested conditions;
or exclusion of the earlier factor prevents disease despite equivalent pathogen exposure.
Microbiome-first is weakened when:
microbiome changes occur only after physiological or pathological decline;
manipulating the proposed microbial state does not alter susceptibility or progression;
or microbial restoration fails to rescue the predicted function.
Environmental-stress-first is weakened when:
decline occurs independently of the registered stress contrast;
the physiological pathway predicted from stress is absent;
or biological initiation clearly precedes the environmental perturbation.
Single-agent models are weakened when:
the agent alone fails to reproduce the observed disease trajectory but specified combinations do.
General multi-pathway models are weakened scientifically when:
they become so permissive that every possible result can be explained after observation.
A framework that accepts every sequence predicts none.
24. Evidence Classification
Claims under this framework should be explicitly labeled.
Current evidence strongly demonstrates
Use only where replicated experimental, pathological, physiological, epidemiological, or equivalent evidence establishes the relevant relationship.
Current evidence suggests
Use where converging evidence supports a relationship but causal direction, generality, or mechanism remains incomplete.
Uncertain
Use where evidence is conflicting, sparse, indirect, or unable to discriminate competing explanations.
Hypothesis
Use for a proposed causal relationship not yet demonstrated.
Prospective prediction
Use only when the predicted outcome is specified before the relevant confirmatory observation.
Retrospective compatibility
Use when an existing observation can be explained by the framework but was not predicted beforehand.
Speculation
Use where the possibility is scientifically conceivable but presently weakly grounded.
These categories should not be silently upgraded.
25. The Relationship to Established Forest Science
The framework does not claim that multifactorial tree decline is a new discovery.
Forest pathology, forest entomology, ecophysiology, disturbance ecology, epidemiology, and tree-mortality science already recognize interactions among host condition, pathogens, insects, drought, climate, soil, and other environmental factors.
Contemporary research describes mortality as an emergent product of interactions among host, biotic agents, and environment. Drought-mortality research has likewise demonstrated the importance of hydraulic dysfunction and interconnected carbon, water, and defense processes. Climate change can alter pathogen biology directly while simultaneously changing host physiology and microbial relationships.
The contribution proposed here is therefore more specific.
It is a protocol for asking:
Which causal sequence occurred in this system, and what evidence discriminates that sequence from plausible alternatives?
Its novelty, if any, must be demonstrated through improved causal discrimination, prediction, or experimental design rather than by redescribing established ecological complexity.
26. Translation Into TSTOEAO
The Swygert Theory Of Everything AO should enter this framework only after the biological variables are defined.
The foundational relation is:
[ V=E\times Y, ]
where V is Value or realized outcome, E is Energy or Opportunity, and Y is Encoded Equilibrium.
The TSTOEAO Empirical Core makes clear that the multiplication sign is not automatically literal scalar multiplication. The minimum empirical interpretation is conditioned expression: available input alone does not determine outcome; independently specified architecture also matters.
A biological translation can therefore be proposed cautiously.
Let:
[ E_t ]
represent a registered biological exposure or available perturbation at time t, such as pathogen inoculum, insect pressure, drought intensity, or another specified input.
Let:
[ Y_t ]
represent an independently measured host–environment state relevant to the registered outcome. Depending on the domain module, this might include genotype, hydraulic state, tissue integrity, defense capacity, root condition, microbial architecture, nutrient status, or other prespecified variables.
Let:
[ V_t ]
represent the registered outcome measured by a fixed receiver: infection probability, lesion expansion, hydraulic loss, mortality, recovery, pathogen load, or another measurable endpoint.
The TSTOEAO claim would not be that all tree decline is reducible to one scalar Y.
It would be that comparable input may produce measurably different outcomes when independently specified Encoded Equilibrium differs.
That is an empirical proposition only if Y is defined before the outcome is known.
27. EC-1: Conditioned Expression in Tree Decline
The relevant Empirical Core proposition is:
EC-1: Conditioned Expression — “Comparable input can produce measurably different outcome when independently specified Encoded Equilibrium differs.”
A forest-pathology test might compare standardized pathogen exposure across trees with independently measured host states.
A qualified prediction would need to specify before outcome:
the exposure;
the relevant component of Encoded Equilibrium;
the predicted direction;
the outcome;
the time window;
the minimum effect of scientific interest;
the receiver;
the null model;
and the forbidden result.
If the proposed host state fails to discriminate outcomes under a qualified test, the local TSTOEAO claim must be weakened or falsified according to the registered rule.
Established biological models must remain explicit comparators.
If conventional plant physiology predicts the result equally well or better and TSTOEAO adds no locked restriction or novel discrimination, the result may be compatible with TSTOEAO without establishing scientific distinctness.
28. EC-2: Channel-Selective Expression
EC-2: Channel-Selective Expression proposes that a change in Encoded Equilibrium can alter which registered routes are admissible, their weights or transformations, what a fixed receiver records, or where preregistered cost becomes expressed; model-defined change alone is insufficient, because at least one independently measured route-specific quantity or receiver-accessible outcome must change.
This maps naturally onto alternative biological routes.
A wound may open a route that intact bark excludes.
Root damage may alter hydraulic routing.
A vector may create a transmission route unavailable without it.
Host genotype may alter pathogen establishment.
Microbial state may alter resource or signaling relationships.
But the biological mechanism must be demonstrated.
Calling a process “channel-selective” after observing the outcome is not enough.
The route must be registered and measured.
29. EC-3: Structured Response
The tree can also be treated as a system responding to declared gradients through declared boundaries.
A drought gradient acts through hydraulic and physiological boundaries.
Herbivory acts through tissue, defense, carbon, and repair boundaries.
Pathogen invasion acts through anatomical and immune/defense boundaries.
The system may respond through successful correction, delayed correction, failed correction, persistence, reorganization, oscillation, temporary compensation, path-dependent transition, or collapse.
The important TSTOEAO requirement is cost.
A defense response may consume carbon.
Stomatal closure may conserve water while reducing carbon gain.
Compartmentalization may limit pathogen spread while sacrificing tissue.
Root loss may shift resources.
A microbial reorganization may compensate or further destabilize function.
These interpretations become scientific only when the predicted cost, recipient, direction, and measurement are specified before the outcome.
30. EC-4: Recursive Boundary Construction
The most direct bridge between TSTOEAO and causal-history analysis may be:
EC-4: Recursive Boundary Construction — “A realized outcome, correction, cost, feedback record, or preserved memory from cycle n causally contributes to the Encoded Equilibrium governing cycle n+1.”
Tree decline is inherently historical.
Defoliation this year can alter stored carbon next year.
A wound can alter future tissue structure.
Drought can change subsequent hydraulic capacity.
Prior infection can alter defense.
Microbial reorganization may change later interactions.
Root loss can alter later water acquisition.
Successful recovery may also change future state.
In shorthand:
[ V_n \rightarrow Y_{n+1}. ]
But recursion must not be declared merely because two consecutive states differ.
The empirical core requires a measurable causal pathway preserving and transferring the earlier outcome into later architecture.
That requirement makes recursive boundary construction experimentally meaningful.
31. TSTOEAO Must Not Determine the Biological Answer in Advance
The biological evidence must be allowed to defeat the preferred theoretical interpretation.
If the evidence shows:
[ Fungus \rightarrow Tree\ Decline \rightarrow Insect\ Colonization, ]
the framework must accept fungus-first.
If it shows:
[ Insect \rightarrow Fungus, ]
it must accept insect-first.
If it shows:
[ Drought \rightarrow Hydraulic\ Failure ]
with no biologically important pathogen or insect role, it must accept that pathway.
If several factors are jointly necessary, the model must accept interaction.
If no proposed sequence survives testing, the model must preserve the unknown.
The governing scientific principle is:
“A theory cannot claim courage before an experiment and become metaphor after the result.”
TSTOEAO is therefore useful here only to the extent that it increases the precision of prospective biological questions.
It cannot be used to force biological evidence into a predetermined sequence.
32. A Minimum Preregistered Tree-Decline Record
Before confirmatory outcomes are examined, a causal-reconstruction study should declare where possible:
host species and population;
system boundary;
unit of inference;
initial state;
candidate causal agents;
competing causal graphs;
proposed temporal sequences;
exposure definitions;
physiological variables;
microbial compartments;
environmental covariates;
route definitions;
outcome measures;
receivers or instruments;
sampling schedule;
confounders;
missing-data assumptions;
intervention or observational estimand;
minimum effect of scientific interest;
statistical model;
alternative models;
support conditions;
weakening conditions;
falsification conditions;
treatment of unmodeled causes.
This prevents the final diseased tree from rewriting the original hypothesis.
33. Research Program
The framework suggests a staged research program.
Stage I: Retrospective reconstruction
Use existing forest plots, pathology records, weather data, historical collections, tree rings, remote sensing, and archived biological material to identify candidate sequences.
This stage generates hypotheses.
It should not be confused with prospective confirmation.
Stage II: Prospective observation
Enroll healthy trees and measure them repeatedly before decline.
Identify temporal predictors without changing exposures.
Stage III: Controlled manipulation
Test individual factors, combinations, and exposure order.
Stage IV: Mechanistic intervention
Exclude proposed initiators, restore proposed lost functions, or interrupt proposed feedback edges.
Stage V: Independent replication
Repeat successful discriminations in different populations, laboratories, forests, species, or environmental regimes.
Stage VI: Predictive deployment
Use the resulting model to predict transitions in trees not used to construct it.
The strongest success would be prediction before visible decline.
34. Predictions
The framework makes several prospective predictions.
Prediction 1
For at least some currently recognized multi-agent decline systems, prospective observation will identify biologically meaningful events that consistently precede the agent most conspicuous at terminal diagnosis.
Prediction 2
The identity of that earlier event will not be universally insect-based, fungal, bacterial, viral, microbial, or environmental.
Prediction 3
Exposure order will materially affect outcome in a subset of multi-agent systems.
Prediction 4
The same biological agent will occupy different causal roles under different host or environmental states.
Prediction 5
Some organisms abundant in terminal decline will contribute little to initiation, while some early events that are difficult to detect at diagnosis will materially alter later transition probability.
Prediction 6
Physiological measurements will often provide causal bridges between environmental and biological stages that categorical diagnosis alone misses.
Prediction 7
Microbiome changes will separate into causal, mediating, compensatory, consequential, and incidental classes rather than forming one uniform “dysbiosis” category.
Prediction 8
Models incorporating time-resolved state transitions will discriminate some mortality trajectories better than models based solely on terminal agent presence.
These predictions require operational definitions before qualified testing.
35. What Would Weaken the Framework?
Evidence would weaken the practical value of temporal causal reconstruction if, across well-designed prospective studies:
terminal diagnoses already captured the complete causally relevant sequence in nearly all tested systems;
earlier detectable states failed to improve causal discrimination or prediction;
exposure order had no meaningful effect where the framework predicted sequence dependence;
multi-agent histories added complexity without improving prediction, intervention, or explanation;
or candidate causal graphs repeatedly proved observationally indistinguishable and experimentally irrelevant.
Evidence would weaken a TSTOEAO-specific contribution if conventional causal, physiological, pathological, or ecological models captured the same results equally well or better without any additional preregistered restriction supplied by TSTOEAO.
Compatibility is not distinctness.
36. Scientific Provisionality
Scientific knowledge is provisional, but provisionality does not mean all explanations deserve equal weight.
A demonstrated pathogen relationship should receive more weight than an untested alternative merely because the alternative is conceivable.
A replicated experiment should receive more weight than anecdotal observation.
Prospective longitudinal evidence should generally receive more causal weight than retrospective pattern matching.
Mechanistic intervention should generally receive more causal weight than co-occurrence alone.
At the same time, strong present evidence does not make a model permanently immune to revision.
New diagnostics can reveal previously invisible organisms.
Genomic tools can distinguish strains formerly treated as equivalent.
Remote sensing can reconstruct trajectories that field observers missed.
Historical collections can preserve evidence unavailable to contemporary witnesses.
Physiological instrumentation can identify thresholds invisible from external symptoms.
Future evidence may therefore change current causal interpretations.
The correct scientific position is neither permanent certainty nor permanent doubt.
It is evidence-weighted revisability.
37. Discussion
The development of the present framework began with a narrower observation: in some tree declines, insect activity may precede conspicuous fungal disease.
That proposition remains biologically legitimate.
It is no longer sufficient as a general framework.
The more durable question is not whether insects generally come first.
It is whether forest decline research can distinguish the history of a dying tree from the state in which investigators happen to encounter it.
That question survives every broadening of the causal field.
It applies if the initiating event is an insect.
It applies if the initiating event is a fungus.
It applies if the initiating event is bacterial, viral, environmental, mechanical, physiological, microbial, genetic, or unknown.
It also applies when no unique initiating event exists.
The framework therefore moves from agent priority toward causal reconstruction.
That change preserves the useful observation underlying the earlier insect-precursor hypothesis while making it vulnerable to stronger tests.
The scientific objective should not be to prove that an overlooked precursor always exists.
The objective should be to determine when one exists, what it is, whether it matters, and how confidently its role can be distinguished from competing explanations.
38. Conclusion
A dead or declining tree can contain several true stories at once.
It may be genetically vulnerable.
It may have experienced drought.
Its roots may have been damaged.
Its hydraulic safety margin may have narrowed.
Its microbial partnerships may have changed.
An insect may have wounded it.
That insect may have carried a pathogen.
A fungus may have produced lethal disease.
Additional organisms may have exploited damaged tissues.
Decomposers may have arrived after irreversible decline.
Each observation can be true without each process having the same causal role.
The scientific task is therefore not merely to identify everything present.
It is to reconstruct the transitions.
The central distinction of this paper is:
demonstrating that an organism can cause disease is not the same as demonstrating that it initiated a particular decline sequence.
The converse is also true:
demonstrating that an event occurred first is not the same as demonstrating that it caused the later death.
A rigorous causal history requires chronology, mechanism, counterfactual reasoning, competing hypotheses, intervention where possible, and explicit failure conditions.
The proposed Temporal Causal-Reconstruction Framework therefore asks investigators to replace the single question—
What killed this tree?
—with a structured sequence:
What was the initial state?
What changed first?
Which change materially altered the trajectory?
What routes became available afterward?
What transmitted damage?
What amplified it?
What physiological transitions occurred?
What remained recoverable?
What pushed the tree across an irreversible threshold?
What arrived only after that threshold?
What would have happened if each candidate factor had been absent?
Which competing causal history best survives the evidence?
TSTOEAO may contribute to this work through Encoded Equilibrium, boundary conditions, channel-selective routing, structured correction, cost, and recursive boundary construction, but only under the same constraint imposed on every competing model: its variables must be defined before outcomes, its predictions must be capable of failure, and its interpretation must remain subordinate to biological evidence.
A tree does not owe the investigator a simple cause.
The history must be reconstructed.
And the reconstruction must remain capable of being wrong.
Claim Classification
Conventional knowledge. Tree mortality and forest disease can involve interactions among host condition, pathogens, insects, environmental stress, hydraulic function, carbon dynamics, and other factors.
TSTOEAO interpretation. A tree's measured biological state may be treated as part of the Encoded Equilibrium conditioning how subsequent perturbations become expressed, provided that state is independently defined and measured.
Retrospective compatibility. Existing examples of insect–fungus complexes, drought–insect interactions, host predisposition, pathogen-mediated mortality, and state-dependent disease are compatible with temporal causal reconstruction but do not uniquely validate it or TSTOEAO.
Prospective prediction. Some tree-decline systems should show measurable sequence dependence, state-dependent routing, and causal effects of prior outcomes on subsequent susceptibility.
Speculation. Particular untested microbiome, viral, historical, or hidden-precursor pathways remain hypotheses until independently demonstrated.
Unresolved question. How often does terminal diagnosis fail to identify the event that materially initiated a decline trajectory?
Evidence that would weaken a claim. Qualified prospective and experimental studies showing no relevant temporal, mechanistic, predictive, or intervention advantage over established single-agent or conventional multifactorial models would weaken the corresponding claim.
Selected References and Source Basis
Adams, H. D., et al. (2017). A multi-species synthesis of physiological mechanisms in drought-induced tree mortality. Nature Ecology & Evolution.
Anderegg, W. R. L., et al. (2015). Tree mortality from drought, insects, and their interactions in a changing climate. New Phytologist.
Choat, B., et al. (2018). Triggers of tree mortality under drought. Nature.
Feau, N., Hessenauer, P., Robin, C., & Tanney, J. B. (2026). Emerging tree diseases driven by climate change: A critical perspective on current challenges and future directions. Annual Review of Phytopathology.
McDowell, N. G., et al. (2022). Mechanisms of woody-plant mortality under rising drought, CO₂ and vapour pressure deficit. Nature Reviews Earth & Environment.
Simler-Williamson, A. B., Rizzo, D. M., & Cobb, R. C. (2019). Interacting effects of global change on forest pest and pathogen dynamics. Annual Review of Ecology, Evolution, and Systematics.
Swygert, J. (2025). Arbor Ecology: The Insect-Precursor Series—A TSTOEAO Framework for Catastrophic Tree Collapse.
Swygert, J. (2026). The Forest Before the Blight: A Microbiome-Preconditioning and Insect-First Cascade Hypothesis for the Collapse of the American Chestnut.
Swygert, J. (2026). Insect Initiation and Interacting Pathways in Northeastern Forest Decline: A Hypothesis on Injury, Pathogen Delivery, and Disruption of Root Microbial Partnerships.
Swygert, J. (2026). Rethinking Tree Decline: A Global Multi-Pathway Framework—A Worldwide Proposal for Reconstructing Causal Sequence Among Insects, Fungi, Pathogens, and Environmental Stress.
Swygert, J. (2026). Beyond Single-Pathway Explanations of Tree Decline: An Open Causal Framework for Biological, Environmental, and Interacting Routes.
Swygert, J. (2026). TSTOEAO Empirical Core v1.0.0: Canonical, Version-Controlled Scientific Specification for Conditioned Expression, Channel-Selective Routing, Structured Correction, and Recursive Boundary Construction. TSTOEAO-EC, Version 1.0.0, Candidate Canonical Draft.
