Definition:
Deep processing (also called semantic processing) is the encoding of information by attending to its meaning, associations, context, and personal relevance, as opposed to shallow processing that focuses on surface features (appearance, sound). The concept originates from Craik and Lockhart’s (1972) Levels of Processing framework, which demonstrated that the depth at which information is processed during encoding — not the amount of time spent or the intention to learn — determines how well it is retained.
In-Depth Explanation
Shallow vs. Deep Processing
Processing exists on a continuum from shallow (structural) to deep (semantic):
| Level | Focus | Example (word: “TABLE”) | Retention |
|---|---|---|---|
| Structural | Visual features | “Is it written in capital letters?” | Weakest |
| Phonological | Sound features | “Does it rhyme with ‘fable’?” | Moderate |
| Semantic | Meaning and associations | “Is it a type of furniture?” | Strongest |
The key finding: processing for meaning produces dramatically better recall than processing for surface features, even when study time is identical and when learners are not explicitly trying to memorize.
Mechanisms
Deep processing produces durable memories because it:
- Creates more retrieval routes — semantic processing connects the item to existing knowledge networks in the mental lexicon, creating multiple paths to retrieve it later
- Engages elaboration — asking “what does this mean?” and “how does it relate to what I know?” forces the brain to construct a richer representation
- Produces distinctiveness — items processed deeply become more discriminable in memory, reducing interference
Application to Language Learning
The implications for vocabulary acquisition and SLA are direct:
- Looking up a word and noting its definition is shallow processing → weak retention
- Using the word in a sentence you create is deeper → better retention
- Explaining how the word connects to a concept you already know is deepest → strongest retention
- Sentence mining is inherently deep — finding a word in authentic context and creating a card with personal meaning
- Keyword method (mnemonics) creates vivid associations → deep encoding
- Active recall combines deep encoding with the testing effect for maximum retention
This is why passive review (re-reading vocabulary lists) is ineffective despite feeling productive — it operates at a shallow level where the illusion of knowing substitutes for genuine learning.
Key Researchers
- Fergus Craik and Robert Lockhart — Levels of Processing framework (1972)
- Jan Hulstijn and Batia Laufer — Involvement Load Hypothesis, applying depth of processing to L2 vocabulary