Tasks sharing a resource pool interfere more than tasks using different pools

What it is

When two tasks run at the same time and sit on different resource dimensions — say one uses visual perception, spatial coding, and a manual response, while the other uses auditory perception, verbal coding, and a vocal response — they interfere with each other far less than two tasks that land on the same dimensional combination (both requiring visual scanning plus manual input, for instance). This is the most direct applied conclusion from multiple resource theory: check whether two tasks draw on the same resource before judging whether they can genuinely run in parallel, rather than judging by how hard each one is on its own.

Why it happens

Cross-modal pairing is cheaper because different resource dimensions have relatively independent supply: two tasks on different dimensions run, in effect, on two separate production lines at once, without queuing behind each other. Two tasks on the same dimension have to take turns on the same production line — even if neither task is individually hard, the turn-taking itself slows both down. This is why everyday experience finds "watching the road while listening to spoken navigation" much easier than "watching the road while reading text directions" — the former splits vision and hearing across separate perceptual dimensions, while the latter crams both tasks onto the same visual dimension, forcing them to queue.

Studying it

The standard empirical approach holds each task's single-task difficulty fixed and varies only how their perceptual or response channels are assigned — same-channel versus cross-channel combinations — then compares the resulting dual-task cost. This contrast isolates the independent contribution of "channel overlap" as a variable, and is the key evidence distinguishing multiple resource theory from a blanket "total difficulty explains everything" account.

Where it stops holding

Cross-modal doesn't mean zero interference. As long as two tasks share any single dimension — even with different perceptual channels, if a visually presented text and a spoken utterance both ultimately require converting to language to be understood, they're still competing on the same coding dimension — "cross-modal" hasn't actually been achieved: switching only the perceptual channel while the cognitive coding stays the same yields limited benefit. Also, when either task draws on central coordination or sequencing resources (deciding which to handle first, switching priority frequently), that coordination overhead causes interference on its own, even when the two tasks sit in different resource pools — this cost isn't protected by channel separation. Under high intensity or high-priority conflict (both tasks demanding an immediate response), a serial bottleneck at the core decision stage can still cause interference even across different resource pools, just at a much smaller scale than same-pool interference would produce. This conclusion rests on laboratory dual-task data; in real settings, two tasks that look cross-modal often still overlap at the coding level, so each dimension needs checking individually rather than concluding safety from perceptual channel alone.

Applying it

When designing parallel information streams, don't stop at "one uses the eyes, one uses the ears" and call it safe — keep checking whether both streams require converting to language to be understood (the coding dimension) and whether both require an immediate manual response (the response dimension); the interference-reduction benefit of cross-modal design only shows up when perception, coding, and response are all kept apart as much as possible. In high-priority, frequently-switching scenarios, even with dimensions separated, control how often the two tasks "cut in" on each other, so central coordination overhead doesn't cancel out the cross-modal benefit. Verification: measure each task's single-task baseline, then measure the performance drop when they run together; if the drop is close to zero, dimensional separation is doing its job; if a clear drop remains, check each dimension individually for an overlooked overlap at the coding or central-coordination level.

Related

• Same group: A9.10.1 attention resources aren't one pool — they split into pools by channel, stage, and response · A9.10.3 vision and hearing are separate perceptual resources, and speech and manual response are separate response resources · A9.10.4 multiple resource theory explains why dual-task interference can't be predicted from task difficulty alone · A9.10.5 designing parallel tasks should deliberately assign them to different resource pools to reduce interference
• Nearby: A5.02 divided attention and dual-tasking (interference severity between same-resource tasks, and how automaticity affects resource demand)
• Search terms: cross-modal interference · resource overlap · dual-task cost