We often talk about psychoactive drugs as though they contain psychological adjectives.
Alcohol is disinhibiting. Benzodiazepines are anxiolytic. Stimulants are activating. Cannabis is relaxing. Psychedelics are mind-opening. Antipsychotics are calming. Antidepressants improve mood.
Each description is useful at the population level and incomplete at the level of the person.
A drug enters a system that is already in motion.
The person has a baseline state, a history of prior exposures, a particular receptor architecture, expectations, sleep, stress, metabolism, current goals, learned compensations, and a context. The effective psychological result is the interaction between all of these and the perturbation produced by the drug.
This can be summarized as:
effect = f(drug, baseline, dose, history, context, physiology, time)
The baseline term is the heart of the theory.
Imagine a variable such as arousal. A person begins below the range optimal for sustained work. A stimulant moves them upward and performance improves. Another begins near the optimum and becomes jittery. A third starts high because of stress and sleep deprivation and crosses into panic or perseveration.
The drug is the same. The movement is not.
The same logic applies to emotional intensity. A person overwhelmed by affect may experience dampening as liberation. A person whose emotional range is already restricted may experience the same dampening as emptiness. Reduced threat can release social behavior in one person and remove needed caution in another.
This is why average drug effects can conceal important subgroups. If half a sample improves and half worsens, the mean can suggest little effect even though the treatment is powerful in both directions.
Pharmacokinetics are part of the explanation but not the whole. People differ in absorption, metabolism, distribution, and elimination. Yet even identical exposure does not guarantee identical response. Pharmacodynamic differences arise from receptors, transporters, circuits, adaptation, and baseline computational state.
Expectation and context also belong in the function. A pharmacological effect is not contaminated because a person’s interpretation contributes to it. The brain’s response to a drug includes what the person believes is happening, what environment they are in, and what behavior becomes possible there. A sedative taken alone before sleep is not psychologically identical to the same exposure taken at a party.
This complicates the notion of a “drug effect”, but it makes it more realistic.
History matters especially. Prior exposure can produce tolerance, sensitization, conditioned responses, and learned associations. The first dose and the hundredth dose do not act on the same system. Withdrawal states alter baseline. A person who uses a drug to correct a withdrawal-induced deficit may report feeling “normal”, even though normality now depends on the drug.
The drug vector itself is therefore state-dependent.
A simple model might represent the post-drug state as x' = x + δ. Real pharmacology is closer to x' = G(x,u,h,t), where the perturbation depends on the starting state, intervention, history, and time.
The clinical consequence is precision psychopharmacology. If baseline measures tell us something about x, treatment can be conditioned on them. This does not require perfect mechanistic knowledge. A useful predictor earns its place if it changes the expected ranking of interventions.
The logical principle is straightforward. If one subgroup does better with treatment A and another with treatment B, a policy that can identify the subgroups should outperform a policy that gives everyone the same thing.
The scientific difficulty lies entirely in identifying the subgroups reliably.
Personality traits may help. So might reward-learning tasks, sleep patterns, autonomic measures, family history, prior treatment response, pharmacogenomics, or biological markers. No source of information deserves privilege in advance.
The baseline law also changes how adverse effects are understood. Emotional blunting, activation, akathisia, sedation, mania, compulsivity, and psychosis are not merely unwanted properties attached to a molecule. They are outcomes generated when a perturbation intersects a vulnerable configuration.
This does not absolve the drug of causality. It makes causality conditional.
The same insight can be applied to enhancement. A drug that improves confidence in a highly inhibited person may push an already dominant person toward recklessness. An intervention that increases openness in a rigid thinker may destabilize someone already prone to unusual beliefs. A focus-enhancing state may benefit scattered cognition and trap a perseverative mind.
There is no trait-targeted pharmacology without baseline assessment.
This principle should become almost boring by the end of the book, because it is the recurring answer to apparently contradictory drug experiences.
Why does the same stimulant calm one person and agitate another? Baseline and network architecture.
Why does cannabis relax one person and provoke panic in another? Baseline, dose, expectation, and context.
Why does an SSRI feel relieving to one person and emotionally flattening to another? Baseline and desired range.
Why does a psychedelic experience become therapeutic insight for one person and destabilizing overbelief for another? Baseline integration, context, and post-state processing.
The law does not predict the outcome by itself. It tells us what information prediction requires.
That is enough to expose a weakness in diagnosis-first prescribing. Diagnostic labels capture important phenotype, but they are often too broad to specify the starting state that matters for a particular perturbation.
The next chapter turns that weakness into a research program.