Philosophy deals with the notion of cause and effect, both identifiable with the latter following the former.

What I have never seen so would like to ask is

How does Philosophy address feedback and feedforward processes in relation to cause and effect ?

53 minutes ago, studiot said:

Philosophy deals with the notion of cause and effect, both identifiable with the latter following the former.

What I have never seen so would like to ask is

How does Philosophy address feedback and feedforward processes in relation to cause and effect ?

Like a traffic jam that starts with a flash of brake lights on a busy road?

Some say this is how our brain deal's with reality.

Correct me if I'm wrong, but isn't this a variation of the tree making a sound in a forest?

17 minutes ago, dimreepr said:

Like a traffic jam that starts with a flash of brake lights on a busy road?

Not quite what I was thinking but yes another example to be explained.

18 minutes ago, dimreepr said:

Correct me if I'm wrong, but isn't this a variation of the tree making a sound in a forest?

No I don't think so.

Turn your Fender or Marshal up and place the mike in front of the speaker.

Electronics is full of examples of feedback. However feedback cannot take you back in time. So the first signal round a feedback loop ignores the feedback loop and so that first time around can be very different to the very quickly established stable effect. This known as a transient condition. I'm not a philosopher but this seems relevant to me. It seems to indicate to me that feedback is in its way cause and effect.

A loop that has no cause must be a paradox...

I am of the opinion that causality is a fundamental aspect of reality, that hasn't been codified by Physics yet.
Quite a few things, such as FtL travel, time travel, etc., are 'allowed' by our models, with the final arbitrer being causality violation, but no specific indication of how the principle of causality actually works in such cases.

Feedback and feedforward would then simply be the non-linear properties of causality.
Not so much a Philosophy problem ( although I'd like to hear their take on the issue ), but a Physics issue that needs overdue addressing.

53 minutes ago, MigL said:

Quite a few things, such as FtL travel, time travel, etc., are 'allowed' by our models, with the final arbitrer being causality violation, but no specific indication of how the principle of causality actually works in such cases

It seems to me that this is precisely what GR does - it tells you mathematically how the lightcones at each event in these circumstances are related to one another. That’s causal structure.

The philosophical question is rather why not all such situations seem to be realised in nature, even though they are valid solutions to the relevant equations.

1 hour ago, Markus Hanke said:

It seems to me that this is precisely what GR does ... That’s causal structure.

Sure some aspects of GR support this causal structure.
But GR also allows for the Alcubierre metric and Closed Timelike Curves, both of which allow for causality violations.
IOW, there is no fundamental principle which clearly defines where causality is violated, and why.

I do know S Hawking proposed a Chronology Protection Conjecture based on the thinking that causality is an emergent property, but this was essentially a convenient opinion to prevent paradoxial situations arising from his work on Black Holes.

6 hours ago, Markus Hanke said:

It seems to me that this is precisely what GR does - it tells you mathematically how the lightcones at each event in these circumstances are related to one another. That’s causal structure.

The philosophical question is rather why not all such situations seem to be realised in nature, even though they are valid solutions to the relevant equations.

Two things seem to enforce causality within relativity:

(1): The 3+1 metric signature ensures that the solution of the electromagnetic wave equation allows only spacetime regions on and inside the lightcone to be affected by a past location. Unlike a 2+2 or 4+0 metric signature, a 3+1 metric signature distinguishes space and time, with space always being the 3 and time always being the 1, in accordance with the nature of lightcones.

(2): Future-directed timelike or lightlike energy-momentum vectors always sum to future-directed timelike or lightlike energy-momentum vectors. The presence of any energy-momentum vectors not so restricted will allow arbitrarily directed energy-momentum vectors to exist.

In addition to the above, transition probabilities are not time-reversible with the resulting cause and effect being statistically valid only in the future direction.

13 hours ago, MigL said:

Sure some aspects of GR support this causal structure.
But GR also allows for the Alcubierre metric and Closed Timelike Curves, both of which allow for causality violations

This is true. However, I would question whether the initial/boundary conditions required to arrive at these solutions are actually physically realisable - the standard Alcubierre metric would require exotic matter, and the Kerr metric - while being a very useful textbook case - requires asymptotic flatness amongst other things, and is unstable under perturbations, so arguably it wouldn’t arise in the real world, except as an approximation in its exterior region.

I would argue that perhaps ensuring that initial/boundary conditions being as physical as possible would avoid most, if not all, such artefacts.

8 hours ago, KJW said:

In addition to the above, transition probabilities are not time-reversible with the resulting cause and effect being statistically valid only in the future direction.

That’s a very good point, I never looked at it that way before.

Maybe 'causality' is the wrong word to use.
We don't make 'measurements' along the time dimension the same way we make measurements along spatial dimensions.
We use sequencing as we can't be at the origin and the end of a time segment simultaneously; we compare the sequencing to another sequence we call a clock, and establish ordering of events.
This sequencing, or ordering, would need entropic consideration, as the ordered state always precedes, and is much less likely, than the many states of disorder ( entropy as the arrow of time ). If I throw the pages of a thousand page book up in the air, there is only one state where the pages are ordered 1 to 1000, but there are 1000! ( factorial, approx 10²⁵⁶⁷ ) states of disorder,
And that's for the ordering of only 1000 events, so it is virtually impossible to go backwards in the sequence.
This also aligns with KJW's statistical consideration of cause and effect.

Or maybe I don't know what I'm talking about, and should leave Philosophy to others ...

Edited November 9Nov 9 by MigL

35 minutes ago, MigL said:

If I throw the pages of a thousand page book up in the air, there is only one state where the pages are ordered 1 to 1000, but there are 1000! ( factorial, approx 10²⁵⁶⁷ ) states of disorder,

Yes and No.

Every single one of those states has exactly the same chance of occurring.

This means that the state you called order is not unique.

In other words you can take any state you like and call it 'order' and make the same statistical argument.

Can we please return to discussing the wider aspect than just relativity. That is why I posted in Pholosophy, not the relativity section.

On 11/8/2025 at 12:47 PM, OldTony said:

Electronics is full of examples of feedback. However feedback cannot take you back in time. So the first signal round a feedback loop ignores the feedback loop and so that first time around can be very different to the very quickly established stable effect. This known as a transient condition. I'm not a philosopher but this seems relevant to me. It seems to indicate to me that feedback is in its way cause and effect.

Thanks for your reply, keep them coming.

In particular you didn't mention feed forward, which includes various arror correcting techniques that are simultaneous.

Edited November 9Nov 9 by studiot

My apologies; I assumed a discussion about causality.

48 minutes ago, MigL said:

My apologies; I assumed a discussion about causality.

No need to apologise, relativistic causality is included.

But for instance I have't seen any biologist responsed about any form of biogenesis or the chicken and egg question.

So please carry on the discussion, I didn't mean to put you off.

On 11/8/2025 at 12:21 PM, studiot said:

How does Philosophy address feedback and feedforward processes in relation to cause and effect ?

I can see the relevance of feedback (both +ve and -ve) but what is the context for 'feedforward' here?

Let me illustrate with a current topic.

Our new short term resident 3i/Atlas is a potential causal factor for a number of possible events.

Its disturbance of the local gravitational field may disturb some small orbiting satellite. If that orbit is inherently stable due to say orbital resonance, then the system will tend to generate reaction forces that oppose and possibly reverse any change in satellite trajectory. These reaction forces would be an example of negative feedback.

A weakly bound satellite whose orbit is merely metastable may find that the disturbance is enough to send it spiralling out of its orbit at an ever increasing rate as the reduction in gravitational attraction with distance is effectively a repulsive force acting in concert with the initial transient. This is a positive feedback response.

So far so good. However, a feedforward response (afaik) implies foreknowledge of a future disturbance from equilibrium which maybe countered by creating an opposing disturbance in advance. So if our best models of solar system mechanics had predicted that 3i/Atlas was going to plough into earth head on, we could potentially launch some countermeasure to deflect the object and reduce its effect. This would be an example of feedforward control.

Nature does do feedforward, but only in living systems afaik. A spooked gazelle may run for its life to avoid becoming dinner. A stranded octopus may 'sprint' over rocks to find a deeper pool. Or is that just an inborn reflex?

The inanimate parts of the universe seem content just to let stuff happen.

2 hours ago, sethoflagos said:

I can see the relevance of feedback (both +ve and -ve) but what is the context for 'feedforward' here?

In feedback techniques the sampling is done at the output, and returned to the input so the correction always lags the output.

In feedforwrd techniques the sampling is done at input and fedforward through a separate path to coincide with the output in a corrective polarity.

Feed Forward Techniques Electronics Today International April 1976 page 68

5 minutes ago, studiot said:

... In feedforwrd techniques the sampling is done at input and fedforward through a separate path to coincide with the output in a corrective polarity.

You're not answering the question I asked.

Feedforward control requires a model (or implicit model) that predicts the effect of a change in input signal on the output.

Since there isn't a hint of retrocausality here, why would a philosopher have any interest in it from a causality perspective?

Chapter and verse at:

Feed forward control

... and Smith Predictor control

where feed forward is a significant component in a system with major industrial relevance (control of distillation towers, paper machines etc)

30 minutes ago, sethoflagos said:

Feedforward control requires a model (or implicit model) that predicts the effect of a change in input signal on the output.

But you are addressing my OP with this statement.

That's discussion for you.

Wouldn't philosophy just try to analyze the core concepts of causality and address any special problems about feedback? I would think they just try to see that the definitions of fb are a good fit with the particular discipline.

An organism interacting with its environment, or a cell interacting with a bodily system, or whatever functional level, would use both positive and negative fb -- often it's about slowing down or accelerating a process, as needed.

Fb is when outputs of a system are routed back as inputs as part of a causal chain that loops. An example would be insulin oscillations. I think the challenges are there for biologists but I don't see much call for a philosopher. Causal chains can loop and it doesn't seem to cause any philosophical crisis AFAICT. Not in biology, anyway, or other fields where the thermodynamic (entropic) arrow is fairly clearly defined. Fundamental physics... maybe a little trickier.

Edited November 10Nov 10 by TheVat

I have to admit to being a little puzzled by the term feed-forward.
Feed-back samples the output and uses that signal to 'correct' the input so as to modulate the output.
Feed-forward uses the input signal to modulate the output directly, but through a separate pathway.
The cause and effect can be the same ( beginning and end ), but the causal path will differ.

I'm watching a video on YouTube with one of the most famous local physicists, and he said:

The coolest questions are philosophical ones, of course, but there's no chance of getting close to them, so we don't ask them, at least if we don't want to make a bad impression. So, we ask questions that are within our reach.

I disagree with the first part, of course, but with the rest, absolutely... ;)

7 hours ago, MigL said:

I have to admit to being a little puzzled by the term feed-forward.
Feed-back samples the output and uses that signal to 'correct' the input so as to modulate the output.
Feed-forward uses the input signal to modulate the output directly, but through a separate pathway.
The cause and effect can be the same ( beginning and end ), but the causal path will differ.

Say we wanted to control the average surface temperature of the planet to 1.5^oC above pre-industrial levels on a long term basis.

Whatever we do, we are denied immediate access to the resulting output that may take many decades to fully present itself. So any attempt at pure feedback control is almost certain to swing chaotically between too-little-too-late and whoops!-too-much. You simply cannot dial in a significant gain value and maintain stability (I have the tee-shirt).

The only real option available is some form of feedforward control of the Smith Predictor type. But that requires a really accurate climate model; full knowledge of all influencing factors; and ruthless elimination of... let's call it 'noise'.

Thinking about it, yes, there are some philosophical issues to address.

On 11/10/2025 at 9:46 AM, Sensei said:

I'm watching a video on YouTube with one of the most famous local physicists, and he said:

I disagree with the first part, of course, but with the rest, absolutely... ;)

The point of philosophy, especially in this context, is to ask and answer what is real.

It's like the same river that can't be crossed twice question, to which the answer is, the bridge hasn't moved and nor has the river...

On 11/10/2025 at 5:49 AM, sethoflagos said:

The only real option available is some form of feedforward control of the Smith Predictor type. But that requires a really accurate climate model; full knowledge of all influencing factors; and ruthless elimination of... let's call it 'noise'.

Thinking about it, yes, there are some philosophical issues to address.

Well, ... yes.
Feedback is self correcting, and require only some limited knowledge of how the system will evolve.
Feedforward only works if you have considerable 'a priori' knowledge of how the system will evolve, so corrections can automatically be made as needed throughout its evolution.
And this method seems not too useful.
If you know enough about the system to predict the corrections that will need to be made so as to restrict deviations and divergences while the system evolves, could you not build the 'corrections' into the initial conditions and eliminate the need for feedforward controls ?

For example ...
Say I'm filling a 10 l reservoir with an inlet flow of 2 l/min and an outlet flow of 1 l/min. and want to prevent overflowing.
After a period of time, the reservoir will be close to overflowing, so a 'feedforward' control will open the outlet to 3 l/min to bring the reservoir down to an acceptable level before resuming the original rate.
This assumes I can calculate when the feedforward controls need to be active to prevent overfilling ( easy calculation ).

But knowing all this, could I just as easily set the initial conditions such that the inlet and outlet flows are equal to prevent overflowing ?
( although if the purpose is to maintain a certain reservoir level, it would work )

Edited November 11Nov 11 by MigL

2 hours ago, MigL said:

Well, ... yes.
Feedback is self correcting, and require only some limited knowledge of how the system will evolve.
Feedforward only works if you have considerable 'a priori' knowledge of how the system will evolve, so corrections can automatically be made as needed throughout its evolution.
And this method seems not too useful.
If you know enough about the system to predict the corrections that will need to be made so as to restrict deviations and divergences while the system evolves, could you not build the 'corrections' into the initial conditions and eliminate the need for feedforward controls ?

For example ...
Say I'm filling a 10 l reservoir with an inlet flow of 2 l/min and an outlet flow of 1 l/min. and want to prevent overflowing.
After a period of time, the reservoir will be close to overflowing, so a 'feedforward' control will open the outlet to 3 l/min to bring the reservoir down to an acceptable level before resuming the original rate.
This assumes I can calculate when the feedforward controls need to be active to prevent overfilling ( easy calculation ).

But knowing all this, could I just as easily set the initial conditions such that the inlet and outlet flows are equal to prevent overflowing ?
( although if the purpose is to maintain a certain reservoir level, it would work )

Sort of. But let's get real.

Your reservoir is also supplied by groundwater springs and rainfall runoff from the surrounding hills. Consumer demand is variable, subject to diurnal and various other cycles. The reservoir is 400 miles away and your level readings are supplied by some old bloke who plods up there and back each morning, and sends you the results by carrier pigeon.