10. Causal influence, surviving information, and recoverability
----------------------------------------------------------------------------------------------------
Suppose A causes B, B causes C, and C causes D. By the time D exists, A may be entirely gone, and
there may be no recoverable record from which any observer could reconstruct it. D can nevertheless
be different because A occurred. Three notions must therefore be kept apart:
causal influence ≠ surviving information ≠ recoverability
Information loss does not amount to causal loss. Burn a letter and its text may become
unrecoverable; but if someone read it first and made a decision because of it, the vanished
information has already altered subsequent reality. Eventually even the reader’s memory may fade
while the consequences of the decision continue to propagate. CTGP is consistent with this: it
entails that the universe’s accessible causal archive is finite and dynamically evolving, shaped by
record formation, thermodynamic degradation, and erasure, without implying that every past truth is
stored in a record.
In a metaphorical sense, reality retains the consequences of its past without containing a memory of
it. “Retains” here means is causally downstream of, not stores a representation of. Reality does not
represent; agents and other information-bearing systems do. The world can be truth-bearing, through
its past-tensed properties, without being representation-bearing.
----------------------------------------------------------------------------------------------------
11. Everything causally active from the past is present now
----------------------------------------------------------------------------------------------------
CTGP’s most exposed empirical commitment is generability: the fundamental laws must fix the
transition from each present state to its successors (deterministically or probabilistically)
without reference to future boundary data, final-state constraints, or global self-consistency
conditions. CTGP adds that enlarging the state space to achieve this is legitimate only if the added
variables are physically instantiated on the present hypersurface and have dynamics of their own.
Combined with the doctrine of causal inheritance, generability yields a sharp principle:
Whatever part of history can still affect the future must already be physically instantiated in
the present.
Otherwise a past that no longer exists would have to reach forward and do additional causal work.
Two consequences follow. First, if two present states agreed in every physically instantiated
respect, their futures would be governed identically by the laws, whatever their histories; history
that left no physical trace in the present retains its truth, through past-tensed properties, but
has no further causal leverage. Second, “instantiated in the present” does not mean “decodable by an
observer.” Causal inheritance can be scrambled, distributed, thermodynamically degraded, or
computationally unrecoverable. Something can be causally preserved while being informationally
unrecoverable.
----------------------------------------------------------------------------------------------------
12. Causality is lawful dependence, not foreknowledge
----------------------------------------------------------------------------------------------------
A cause does not need to “know” its effect. A neuron does not represent the decision the organism
will eventually make; a weather system does not contain a forecast of tomorrow’s weather; and under
CTGP the present does not need to contain a pre-existing future in order to generate one. A process
can produce a definite outcome without itself containing foreknowledge of that outcome.
Causality is a relation of lawful dependence, not a relation of foreknowledge.
Several distinctions follow. Predictability and causation are separable: something can be causally
determined without being predictable by any observer, and something can be highly predictable
without the predictor being its cause. Likewise, a determinate outcome is not the same as a
pre-existing outcome. An event at t₂ may not exist at t₁, exist at t₂, and leave consequences at t₃,
without anything at t₁, including the universe itself, possessing a completed representation of t₂.
Three further concepts should not be conflated: prewritten, lawful, and predictable. Something can
be lawful without being prewritten, lawful without being computationally predictable by an observer,
and unpredictable without being causally unstructured. CTGP’s picture is precisely lawful generation
without prewritten specification: order does not require predetermination, and novelty does not
require causelessness.
Coherence can also be retrospective without being prescient. A completed causal process can be
reconstructed afterward as a continuous chain even though nothing within it possessed foreknowledge
of how it would end.
----------------------------------------------------------------------------------------------------
13. A complete causal history without a complete history book
----------------------------------------------------------------------------------------------------
Suppose the history from A to E was entirely lawful, and afterward all evidence distinguishing A
disappears. There can still be a fully definite answer to the question “Why is E this way?” while no
observer at E, and perhaps no possible later observer, can reconstruct that answer. The signal-based
epistemology explicitly allows such structural limits rather than treating all ignorance as a
temporary shortfall of technology. Hence:
• causal completeness does not entail epistemic completeness;
• intelligibility in principle does not entail reconstructibility in principle;
• reality can be lawful and determinate while being informationally lossy.
A further obstacle remains even where information has not been destroyed. Any observer attempting to
model the whole universe is itself part of the universe, so a complete model would have to include
the modeler and its own modeling activity, which raises problems of resources and self-reference.
The “view from nowhere” may therefore be structurally unavailable to any proper part of reality.
Being part of reality is itself an epistemic condition. In a phrase, reality can be fully
determinate while having thrown away the receipt.
====================================================================================================
PART III. Agency and the Representation of the Future
====================================================================================================
----------------------------------------------------------------------------------------------------
14. Agency does not require exemption from physics
----------------------------------------------------------------------------------------------------
CTGP describes conscious agency, where it exists, as an emergent causal process within lawful
physical dynamics rather than as an extra-physical selector of outcomes. This separates two ideas
that are often run together:
agency ≠ exemption from causality
An agent can be physically constituted, causally influenced, internally structured, capable of
deliberation, capable of representing alternatives, and capable of changing its behavior for
reasons, and still be part of the causal chain. If an agent’s reasoning had no causal influence at
all, calling it an agent would be difficult. Causal efficacy is part of what makes agency
intelligible. A hurricane is caused by atmospheric conditions and is still a real process doing
things; a decision can be caused by prior conditions while the person’s reasoning, memories, values,
and actions are themselves links in the chain. “Caused” and “not really doing anything” are not
synonyms.
The signal-based framework makes a related point about creativity. The brain need not manufacture a
finished idea somewhere “in advance” and then merely reveal it to consciousness; creative
construction can occur through ongoing interaction, recombination, feedback, and modification in the
present. Causal systems can generate genuinely new configurations through their own dynamics.
----------------------------------------------------------------------------------------------------
15. Memories and plans point in opposite causal directions
----------------------------------------------------------------------------------------------------
Under CTGP, a memory and a plan are both present states about something that does not exist: a past
that is gone and a future not yet generated. They relate to their targets in opposite ways. A memory
is an effect of what it represents, since yesterday caused the memory of yesterday. A plan is a
cause, or at least a candidate cause, of what it represents, since a plan for tomorrow is among the
things that may generate tomorrow.
A memory is true because of its causal ancestry. A plan becomes true, if it does, because of its
causal descendants.
This yields a non-mystical definition of agency:
An agent is a causal system whose present representations of possible future states participate
in determining which future state actually occurs.
It also explains why prediction and intention are easily confused. A forecast and a plan can look
identical on paper, but they relate to the world differently:
• Prediction: present model → later correspondence. It succeeds by accurately describing what
happens.
• Intention: present model → causal action → later correspondence. It can succeed partly because it
existed.
A weather forecast predicts tomorrow; a human plan can help cause tomorrow. Nothing here involves
retrocausation. It is not the future affecting the present; it is a present representation of a
possible future affecting the present.
----------------------------------------------------------------------------------------------------
16. Possibilities can matter causally without existing
----------------------------------------------------------------------------------------------------
An intelligent system can represent alternatives: “I could do A; I could do B; if A, probably C; if
B, probably D.” Neither imagined future needs to exist, but their present physical representations
do, and comparing them changes what happens next. Physically existing alternative futures are
therefore not required for alternatives to matter causally.
possible future ≠ existing future, while a representation of a possible future is part of
present causal reality
An open future does not imply an informationally empty present. The present can be full of models of
futures that do not exist, including plans, fears, simulations, forecasts, and contingency trees,
and those present representations help determine which later state occurs. Agency is one way reality
becomes causally responsive to possibilities without those possibilities being realities.
The present is thus an unusual junction. It contains causal inheritance from things that no longer
exist and representations of things that do not yet exist, and only the former came from its
referent:
• Memory: past event → present representation.
• Perception: present interaction → present representation.
• Intention: present representation → action → future event.
The same kind of thing, an information-bearing physical state, can play radically different causal
roles depending on its position in the causal network. A photograph of a fire and a plan to start a
fire may share descriptive content while doing opposite causal work.
====================================================================================================