Glad you asked. There could be quite a few. Numerous devices IRL today have things like gesture or voice control along with hand controls- and not just as alternate or back up options.
On the surface hand controls and neural interface seem redundant- but we have to remember the technology side. There are many reasons specific to a given universe why hand controls may be used.
First consider that certain functions might not be practical with the neural link technology. Perhaps certain thoughts are difficult to translate or lack speed or accuracy or some aspect that hand controls are better at for those functions, or perhaps there just isn’t enough processing in the system to handle every single command through the link.
Next consider that depending on how the link works- it may not be able to adequately filter “random thoughts.” You might want something like the self destruct to use a control other than neural link- what if someone has a random suicidal thought or moment of despair akin to standing near a ledge and thinking of dropping off? Perhaps you might want weapons or other potentially dangerous systems to use a manual safety- as your space mech floats along on hour 6 of convoy escort in space and you day dream at shooting passing ships in your wandering mind to pass time, or your squad mate pisses you off and you think “I could just waste them right now…” you might want a hard wired safety gate in case your system- which in some way takes mental impulse and turns it to a reaction- to not turn every impulse into a war crime.
Let’s say your system has safeguards to that. Like it doesn’t take commands off random thoughts but translates impulses sent to the body- so your arm doesn’t punch if you think about punching- it punches when you send or would send the impulse to actually punch. Now your mech doesn’t react to the thought but the intention. Well… two things to consider. Unless something in the system or else wise has effectively paralyzed your body, if you send an impulse to punch, your arm will punch even though it doesn’t need to. In that sense the hand controls may not be even necessary so much as something to restrict your movement to some range the system can make use of and keep your body from flailing around. What happens if you are knocked from the side with nothing to hood in your hands and your arms fly outstretched and you want to punch? You need to send the impulse to punch-
So either the mech would need to move to always match your body position- which is problematic on several levels and likely not desirable, or if the impulse is what is needed, you need to return your arms to your side to punch right? So you’re either wasting time or building complication into the system. If the cockpit is small, you could also be injured. So securing the extremities makes sense and providing contact points like hand and foot controls is one way to give you stability to have your body stay in a ready position to pilot.
Consider also that the system may not be able to EXACTLY translate thought in the sense of the human body. This could be thought of as a more realistic or less advanced neural control system. How? Ok.
The more types of movement the machine can do and the more complex or intricate those movements- the more complicated neural control becomes in theory. Now, humanoid robot- if we program it well, have computer “translators” behind the neural interface, or the machine is mechanically built like a human anatomy (unlikely and probably not optimal..) let’s say you think about stepping and the machine takes the same impulses you’d use to walk and walks the same. You think of catching a ball and it does the same motor control as your arm etc. but… no part of the human brain is wired to fire our built in lasers or head mounted Vulcan cannons or fire out thrusters or adjust out built in aero foils for mid air maneuvering is it? So in this “direct translation” you can’t get a motor impulse from the brain for thrusters. We don’t have them. You could do things like-
If the thrusters were only for jumping high, you could have a computer fire and control the thrusters automatically based on the intensity of the jump you will yourself to do- so if you think about jumping as high as you can, the thrusters fire and the machine jumps as high as a skyscraper. This would take quite a bit of training and practice to be able to judge jumps and control it and it still wouldn’t allow easy or intuitive flight on impulse, a pilot would need to train to “fly” the way the machine flys, out of the machine in some special rig so they know how to move their body in a way the machine can use. So you’re looking at a fork in design philosophy between a machine that the pilot must adapt to or a machine that is designed to adapt to the pilot. Generally in the real world, between people of relatively equal potential and experience, a machine that is more intuitive and easier to use offers better performance. You generally want to make the machine easy to use as practical
So we are back to a place where to operate things like thrusters and such via direct link you either need some potentially clugey system or you need to translate “imagination” to reality while somehow filtering for intent and with sufficient latency to allow the pilot time to form a conscious decision. We often think or envision actions before we do them, and many cognitive theories hold we rationalize after or have some split second where we decide wether to do or merely think. A direct numeral link would in theory need some delay to allow the pilot to come to resolution.
The three biggest draws to a neural link in theory are:
1. Fast response
2. Allows complicated controls that would be cumbersome and almost impossible for a human to work with hand controls.
3. May reduce operation specific training as the machine simply requires thinking and some procedures. You can focus on tactics and strategy over a “user manual.”
A system with high latency can lose the advantage in speed
Which is a major potential draw for those control schemes, or fall so far behind in speed or ease of use as to make it a liability compared to other war machines.
we run into another problem- where such sufficient technology exists as to create a system capable of complex neural control let alone fast and efficient neural control, and with computers to handle all the “backend” like power management, cooling, translation of movement for the frame vs. The body, attitude control, FCS, coms, etc etc. and can interpret intent and what is appropriate or advisable etc…. We have basically created an AI fighting machine with more steps. Modern high speed aircraft often have flight envelopes that are insane. Look at places from WW2 and before. They tend to be “plane shaped.” Huge wings that go straight out, big cylindrical bodies, so on. Not the best aerodynamic profiles for speed and range and maneuverability or stealth- but… stable. Relatively easy to fly.
Modern planes with their often almost non existent seeming wings- sharp delta (“V”) shaped wings and tiny stabilizers or none etc- their complex yaw controls and such- they generally cannot be flown by a human. Keeping them in the air with such unstable profiles requires constant attention and adjustment from lots of tiny surfaces and very precise and fast control. Computers handle a lot of that and the pilot mostly controls the air speed and gives general direction and lets the plane figure out how to do that and stay in the sky- though our current systems will often allow pilots to do things that can make the plane not be able to stay in the sky, a limitation of technology and a need to balance wether the pilot or the machine is ultimately trusted to decide what a situation calls for and the risk they need to take against the risks present.
Likewise, a machine as complex as many mecha would be need that same level or greater in computer assistance most likely- but if directly thought controlled in the most literal sense they also need the predictive and awareness capabilities discussed above. At that point the pilot isn’t doing much except moving some sticks around. Which could in some cases go back to the “human and hand controls as a safety” reasoning- the concept of letting machines control weapons and letting machines fight amongst humans without a human to act as the “final word” isn’t one most people on average are comfortable with, so in these societies it may simply be that the pilot serves more ad a nanny than a pilot- but we quickly approach a point as the technology gets sophisticated enough for direct neural control where the logic of putting a human inside the machine makes less and less sense with the bodies weaknesses to G forces and deltaV, the need for life support, human frailties like needing to sleep..
But if the pilot is just a safety- why not drone control like we often use today with sophisticated computer aided war machines? Why not have all your pilots on a well defended and or far away place controlling robots so you get rid of the human bulk and problems of varying a squishy person around and don’t risk pilots deaths in the field? Perhaps electronic warfare makes that ill advised? Jamming or hacking? Perhaps the human and the hand controls offer some level of redundancy or hardening. Maybe such neural networks are harder to hack and allow systems to be designed in isolated manners so that your control systems are not on a network or networks are partitioned etc. and ultimately for cyber warfare or mechanical failure the hand controls offer the ability to at least have some function using hard controls?
We haven’t talked about a big one. We touched the complexities of neural interface in theory, a bit. But from the machine side. What about the pilot? It is common in such cases that it is said or alluded to that pilots experience some form of “strain” from controlling such machines. What if the limitation is the pilot? What if there is only so much a brain/mind can control at once or at all, and systems are offloaded to manual controls that are either considered too large a burden for a pilot to bear or to free up bandwidth to allow control of multiple systems that are more beneficial to neural link? Perhaps and in part dictated by the mech design etc- arms are simpler to move with controls than thrusters or weapons? Mech designs, mobility, how “humanoid” they are and the precision and complexity of things like their manipulator or “hand” movements and range of motions of their joints vary. One mech may only be able to raise their arms “up and down” and perhaps some level of “sideways”
so they are easily mapped to hand controls but would be demanding as neural controls or eliminating them as neural controls frees up room for other abilities or makes the system faster etc. speaking of mapping… hand controls could be mapped, various movements and compound movements etc. could be mapped like videogame combos” and some hand controls coukd even be contextual- working differently based on the situation or various factors. This mapping can be mechanical or mental or both. So say that thinking about moving the controls a certain way signals the desire to use some function the human brain has no natural motor impulse for. Again- it’s all very specific to the technology and philosophy and circumstances of a world and society and a specific mech etc. what is the mech used for? Is it designed for a specialized type of combat or combat role or other task?
So indeed- if we are being honest it may often be largely stylistic or nonsensical to have hand controls in a neural linked mech, but it also could make a lot of sense. Many legendary mech designers actually place(d) a lot of thought into their designs- sometimes because the intention was to manufacture toys and the design needed to be something that could be built and actually work and move and possibly transform or accept a pilot figure etc. sometimes because they have backgrounds or interests in mechanical or industrial design or some specific facet and really thought out the way the machine would work if some theoretical technology existed to allow it.
A quote I’ve held on to for many years and attributed in my hearing to Quentin Tarantino is essentially that there are seldom plot holes so much as an audience that lacks imagination. It’s a philosophical argument wether one thinks the creator should spoon feed the audience, at least suspend disbelief with a tight narrative that doesn’t leave an audience potentially questioning or confused; or wether the film maker doesn’t need to show or explain everything in a way that is immediately understandable and the audience can decide to just trust the narrative or determine for themselves how a character gif from A to B or so forth. If we want to be “smarter” than the entertainment we are watching we likely will feel that way.
On the surface hand controls and neural interface seem redundant- but we have to remember the technology side. There are many reasons specific to a given universe why hand controls may be used.
First consider that certain functions might not be practical with the neural link technology. Perhaps certain thoughts are difficult to translate or lack speed or accuracy or some aspect that hand controls are better at for those functions, or perhaps there just isn’t enough processing in the system to handle every single command through the link.
Consider also that the system may not be able to EXACTLY translate thought in the sense of the human body. This could be thought of as a more realistic or less advanced neural control system. How? Ok.
The three biggest draws to a neural link in theory are:
1. Fast response
2. Allows complicated controls that would be cumbersome and almost impossible for a human to work with hand controls.
3. May reduce operation specific training as the machine simply requires thinking and some procedures. You can focus on tactics and strategy over a “user manual.”
A system with high latency can lose the advantage in speed
we run into another problem- where such sufficient technology exists as to create a system capable of complex neural control let alone fast and efficient neural control, and with computers to handle all the “backend” like power management, cooling, translation of movement for the frame vs. The body, attitude control, FCS, coms, etc etc. and can interpret intent and what is appropriate or advisable etc…. We have basically created an AI fighting machine with more steps. Modern high speed aircraft often have flight envelopes that are insane. Look at places from WW2 and before. They tend to be “plane shaped.” Huge wings that go straight out, big cylindrical bodies, so on. Not the best aerodynamic profiles for speed and range and maneuverability or stealth- but… stable. Relatively easy to fly.