Yes — that is a serious epistemological problem, and there isn't a completely neutral standpoint from which to settle it.

The key issue is that “normal” and “altered” are not themselves epistemic categories. They are partly statistical and social categories: a state is called normal because it is the state people tend to occupy, communicate in, and coordinate around. That doesn't automatically mean its representations are more truthful.

From an altered state, your point can become very sharp:

If my ordinary state systematically filters or distorts experience, then using ordinary-state judgments to declare the altered state “unreliable” risks circularity.

But there are ways to distinguish “this state feels extraordinarily meaningful” from “this state gives me reliable information about the world.”

The strongest criterion is cross-state convergence

Suppose:

• In ordinary consciousness, I perceive a table.
• In an altered state, I perceive the table as radically different, perhaps even as something that ordinary consciousness fails to recognize.
• When I return to ordinary consciousness, I can repeatedly inspect the table.
• Other people, independently, can inspect it too.
• We can manipulate it, measure it, predict its behavior, photograph it, etc.

The ordinary state therefore has something the altered state may lack: stable intersubjective and behavioral constraints.

It's not simply:

“Normal consciousness says normal consciousness is trustworthy.”

It's closer to:

“Representations generated in this state continue to predict and coordinate with observations generated across many people, times, and conditions.”

That's a much stronger argument.

But your objection survives

Because an altered state could say:

“Of course you all agree. You're all trapped in the same perceptual framework.”

And logically, we can't simply rule that out.

In fact, we can't prove from inside any cognitive state that the state isn't radically distorting reality. This is closely related to skepticism, dreams, hallucination, simulation hypotheses, and the problem of induction.

So epistemology generally has to give up on absolute certainty and ask instead:

Which mode of cognition produces beliefs that are most robust under attempts to falsify them?

That gives ordinary consciousness a pragmatic advantage, rather than an unquestionable metaphysical privilege.

Altered states can actually be epistemically valuable

This is an important distinction.

An altered state might be unreliable at telling you:

“There is literally a cosmic entity communicating this proposition to me.”

while being extremely valuable at revealing:

“My ordinary assumptions about myself, time, agency, emotion, or perception are not inevitable.”

In other words, an altered state can provide epistemic evidence about the structure and contingency of your ordinary experience, without necessarily providing reliable evidence about whatever external metaphysical reality supposedly produced the experience.

That's a subtle but important middle position.

And “gibberish” may be the wrong category

Imagine that ordinary consciousness imposes a particular syntax on experience: objects, persistent selves, linear time, discrete boundaries, ordinary causality.

An altered state might weaken some of those organizing principles.

From within ordinary consciousness, the resulting experience can look like nonsense.

But that doesn't establish that it is nonsense. It could instead be:

1. noise — genuinely disorganized information;
2. a different representational system — meaningful but difficult to translate;
3. a mixture — genuine insight embedded in hallucination/confabulation;
4. metacognitive information — information primarily about how the mind constructs experience;
5. an experience whose phenomenal meaning exceeds its propositional content.

The fourth possibility is particularly interesting. An altered state may tell you something true without giving you a true sentence about the external world.

For example, the experience:

“The self isn't as solid as I normally experience it.”

might be epistemically significant even if the accompanying interpretation—

“Therefore the universe is literally one conscious organism”

—is unjustified.

There is an asymmetry worth noticing

Normal consciousness has a huge empirical advantage because it is self-correcting across people and time.

But altered consciousness can expose something that ordinary consciousness tends to conceal:

the fact that its own apparent obviousness is itself a feature of a particular state.

That's philosophically powerful.

You could put the problem like this:

Why privilege the consensus of one cognitive regime over the testimony of another cognitive regime?

The answer cannot simply be “because the first is normal.”

A better answer is:

Because the first regime has demonstrated greater reliability through cross-checking, prediction, reproducibility, intersubjective agreement, and successful intervention — while remaining open to correction.

But importantly, that is fallibilism, not certainty.

And there's a fascinating recursive consequence: the very ordinary state that we use to evaluate altered states is itself something we can investigate while in altered states. We can ask what ordinary consciousness systematically excludes, what it constructs, and which of its assumptions survive transitions between states.

So I think your intuition points toward a genuinely deep epistemological principle:

No state of consciousness gets epistemic authority merely by declaring itself the default. Authority has to be earned by the stability, predictive power, and cross-state/cross-person robustness of the knowledge it produces.

That still leaves open the possibility that both ordinary and altered consciousness are partial perspectives on something neither has direct access to.