Learning is traditionally defined as a relatively permanent modification of behavior as a result of experience. In biological systems with a nervous system, several forms of learning are distinguished: habituation (decreased response to repeated harmless stimuli), sensitization (increased response after harmful stimuli), classical conditioning (association between two stimuli: Pavlov), operant conditioning (learning the consequences of one's actions), complex cognitive learning. The simplest forms (habituation and sensitization) are documented even in organisms without a nervous system (Paramecium, Physarum polycephalum): the debate is whether plants surpass this basic level toward more complex forms. Research by Monica Gagliano and other scientists is exploring this frontier.

Pavlovian Conditioning in Plants: The Controversial Experiment

Monica Gagliano (2016, Scientific Reports) published an experiment claiming the first demonstration of classical conditioning (similar to Pavlovian conditioning) in a plant, the pea (Pisum sativum). The experiment: peas are exposed to an air stream (neutral stimulus) coming from the same side where light arrives (unconditioned stimulus for phototropism). After the conditioning period, the air stream is presented only from the opposite side of the light. The hypothesis: if the pea has associated the air stream with light, it will grow toward the side of the air stream (even in the absence of light). The result: conditioned plants tended to grow toward the air stream even without light, unlike non-conditioned plants. The controversy: the experiment has been partially replicated but with inconsistent results. The methodology has been criticized (small sample size, difficulty controlling confounding variables). It is not yet considered a consolidated result by mainstream botany. Gagliano's response: the results are consistent with what is observed in organisms without neurons (Physarum). The burden of proof to exclude associative learning in plants rests with those who contest it, not only with those who propose it.

Anticipatory Adaptation: Proactive Behavior in Plants

A more documented and less controversial form of learning is anticipatory adaptation: the ability of plants to modify present behavior in response to predictive signals of future conditions. The so-called "clock behavior" of plants: many plants prepare anti-herbivore defenses during the hours of the day when herbivores are most active, even in the absence of herbivores themselves. Tobacco seedlings produce defensive phenolic compounds with a peak in the early afternoon (when Manduca sexta caterpillars are most active) and do so in advance of the caterpillars' arrival: a programmed behavior that uses the circadian clock as a predictor of risk. Roots and water: experiments by Gagliano et al. (2012, Trends in Plant Science) and other researchers show that roots of various species are able to "direct" growth toward water sources before even reaching the wet zone, detecting acoustic signals (200 Hz vibration produced by flowing water) at a distance. If confirmed, this would be an example of prospective behavior based on non-chemical predictive signals.

Root Optimizing Behavior: An Intelligence of Resource Allocation

One of the most compelling arguments for a form of learning/intelligence in plants is the optimizing behavior of roots in competition for resources. In conditions of competition with roots of other plants: root allocation changes in an optimizing manner in response to the presence of neighbors. A study by Semchenko et al. (2007, Proceedings of the Royal Society B) showed that roots of some grass species distinguish between their own roots and those of other plants of the same species (kin recognition) and reduce inter-individual competition (producing fewer roots overlapping with those of their "relatives") while increasing intraspecific competition toward plants of other species (producing more roots overlapping). This differentiated behavior (cooperate with relatives, compete with strangers) requires the ability to distinguish between "self" and "other" and to modulate behavior based on this distinction: a form of elementary cognition that many scientists consider remarkable in organisms without a nervous system.

Physarum polycephalum: Intelligence Without Neurons as a Model

To understand plant intelligence, it is useful to look at Physarum polycephalum (the "blob" or slime mold): an acellular eukaryote (a single giant cell with many nuclei) capable of extraordinarily intelligent behaviors despite the complete absence of neurons or nervous structures. Physarum solves mazes (always finds the shortest path between two food sources in a maze), creates efficient networks by optimizing nutrient transport through its network of tubes (the Physarum transport network in laboratory experiments spontaneously replicates efficient networks like Tokyo's subway), shows habituation learning (becomes accustomed to harmless adverse chemical stimuli and ignores them), has distributed memory (the "memory" of food source locations is encoded in the structure of the tube network). Physarum demonstrates that cognition and complex problem-solving do not require neurons: they can emerge from networks of chemical and physical communication even in single-celled systems. This provides a framework for understanding how plants, with their vascular networks and signaling systems, can display functionally intelligent behaviors without the biological structures that animals use for intelligence.

Criticisms and Limitations: What We Still Don't Know

Despite interesting evidence, there are important limitations in research on plant learning that require intellectual honesty. Reproducibility: many of the most spectacular experiments on plant learning have not been robustly replicated by independent laboratories. Science requires replicability. Alternative interpretations: many behaviors attributed to learning in plants can be explained by simpler mechanisms (biochemical oscillators, gradient sensing) without requiring learning mechanisms analogous to those in animals. Lack of clear molecular mechanisms: for vernalization we have a precise molecular mechanism (the Polycomb/FLC cycle). For Mimosa habituation and pea conditioning, molecular mechanisms have not yet been identified. The most important open question: is there something it is "like" to be a plant? Is there subjectivity, experience? Here science is honest: we don't know. And perhaps we never will with current methodologies. Consciousness is the hard problem of neuroscience even for animals: it is even more difficult for plants.

Ethical Implications: If Plants Learn, Should We Treat Them Differently?

Research on plant learning and intelligence has ethical implications that the scientific and philosophical community is beginning to discuss. If plants "learn" and "remember": should we take this into account in agronomic practices? In herbicide use? In intensive monoculture? In forest cutting? Philosopher Michael Marder (Plant-Thinking: A Philosophy of Vegetal Life, 2013) argues that evidence of psychic life in plants (even if not conscious) requires a reconsideration of environmental ethics. Peter Wohlleben (The Hidden Life of Trees, 2015) brought this discussion to the general public with a very effective popular science approach (and criticized by some for excessive anthropomorphism). The most balanced scientific position: plants have inner life in the functional sense (information integration, adaptive behavior, communication, priming, memory) but there is no evidence of subjective consciousness (qualitative experience of the environment). This distinction is important: we should not treat plants as if they were people, but we can recognize their complexity and capabilities without romanticism and without reductionism. Regenerative agriculture, natural afforestation, permaculture: approaches that respect the complexity of plant ecosystems. Not for sentimentality: for scientific effectiveness.