Manovich

Works

Three books tracking the same argument from interfaces to data.

• The Language of New Media (2001) sets out the five principles — numerical representation, modularity, automation, variability, transcoding — and the database as symbolic form.
• Software Takes Command (2013) argues that there is no digital media, only software, and that the tools structure what gets made with them.
• Cultural Analytics (2020) is the methodological statement for computational analysis of large cultural datasets.

Central ideas

Three, and the first is the one this work has most reason to sit with.

• The database as symbolic form: narrative and database are competing ways of organising the world, and digital media privileges the second. A story asserts one order; a database holds items and permits many, so an order becomes a reading rather than a property of what is recorded.
• Transcoding: a media object has a cultural layer and a computer layer, and the second increasingly reshapes the first. Categories that exist because a system needs them start to look like categories the world has.
• Software studies: the specific tools have specific affordances, and studying them is studying the culture they produce rather than a technical preliminary to it.

Distinctive vocabulary

Three terms used loosely wherever they travel.

• Transcoding: the reshaping of a cultural layer by a computational one. Not format conversion.
• Database, in his usage: a symbolic form, a way of organising the world. Not a storage technology.
• Cultural analytics: computational analysis of large cultural corpora. Not analytics in the business sense and not a way of studying individuals.

Debates and disagreements

The arguments are about what computational analysis can reach.

• Whether analysing a corpus computationally finds meaning or only what is countable. It is the same objection made to distant reading in literary studies, and it has not been answered so much as bounded.
• Whether large visualisations present interpretation as though it were given data. The digital-humanities critique holds that what is displayed is always taken rather than given, and that the aesthetic force of a visualisation can substitute for an argument.
• Whether the database and narrative opposition is too clean: databases carry orders and indexes, narratives carry structures, and the distinction may be a matter of emphasis rather than of form.

Intellectual relationships

He shares Kittler's materialism and refuses his determinism.

• Kittler, in this same lens lineage, studies the material substrate archivally and deterministically; Manovich studies it empirically and declines the determinism, which makes them allies in method and opponents in conclusion.
• McLuhan is the field's founder and the least like him in style, since the whole point of cultural analytics is to replace the aphorism with the measurement.
• Russian formalism and the cinema avant-garde supply the genealogy he traces new media back through.
• Distant reading in literary studies is the parallel programme and inherits the same objections.
• The remediation account is the rival theory of how new media relates to old.

How it changes this reading

Manovich is named in the lineage of the Media-theoretic lens, and the database claim describes this work's own form.

• This work is a database with paths through it: claims, lenses, through-lines and placements, presented so that more than one route exists. On his account that is not a neutral presentation choice — it makes some relations findable and others structurally invisible, and the paths that exist are a property of how the collection was built rather than of the life it concerns.
• Transcoding names the specific hazard: a category introduced because the system needed one can start to read as a category the subject has. The distinction between the two layers is exactly what this work has to keep visible about itself.
• His own method's limit applies here too. Counting and linking find what is countable and linkable, so the shape of what this work can surface is bounded by the shape of how it was encoded.
• It is not evidence about how any record was produced or about anyone's use of any system.

Useful comparisons

The lens names four figures with four methods.

• Marshall McLuhan: media as extensions of human faculties, aphoristic and humanist.
• Friedrich Kittler: media as material conditions, archival and anti-hermeneutic.
• Lev Manovich: software and data, empirical and taxonomic.
• Walter Ong: orality and literacy, and the claim that writing restructures consciousness.

Where the ideas meet

None of the four treats a medium as a neutral pipe.

• None treats a technology of inscription as indifferent to what passes through it.
• All four hold that a change of medium changes what is possible rather than only what is convenient.
• Each treats the material conditions of recording as part of the analysis.

Where they part

Shared materialism does not make one programme.

• Kittler's determinism is stated flatly and left unqualified; Manovich's work is empirical and declines it, so the two are allies in what they look at and opponents in what they conclude from it.
• McLuhan's method is the aphorism and Manovich's is the measurement, which are not two styles of one approach but two accounts of what evidence would be.
• Ong's claim is about consciousness and Manovich's is about form; neither supports the other and citing them together for the general proposition that media matter loses both.

Limits

• Cultural analytics is a method for corpora and says nothing about individuals, which is the scale at which it is most often misapplied.
• The five principles are a taxonomy of new media and do not explain why any particular object takes the form it does.

Criticisms

The objections are methodological and mostly conceded to be open.

• Computational analysis finds what is countable, and whether that reaches meaning is the field's standing unresolved question.
• Datasets and metrics carry the biases of their construction, and large visualisations can present a taken measurement as a given fact.
• The work depends on platform data whose availability and shape are set by companies rather than by researchers.
• The five principles describe rather than explain, which limits what they can be asked to do.
• The database and narrative opposition is arguably too clean to survive contact with either form.

Common misreadings

Four uses this placement does not license.

• Applying corpus-scale methods or conclusions to an individual.
• Treating a visualisation as a finding rather than as a display of a measurement someone chose.
• Using transcoding loosely for any format change.
• Citing the database and narrative opposition as a settled distinction rather than as a contested one.

What remains outside this idea

• It cannot establish anything about how a particular record was made or about who made it.
• It cannot convert a claim about form into a fact about a life.