Comparative deletion and comparative subdeletion in translation concern the meanings hidden inside sentences such as “Lina is taller than Omar is” and “Lina has more books than Omar has magazines.” The first compares heights without saying the adjective twice; the second compares the number of one kind of object with the number of another kind. Their English surfaces look economical, but the reader must reconstruct a standard, a measurement dimension and, often, an unspoken degree expression. A translation can preserve every visible noun and still compare the wrong quantities.
Searches for comparative deletion, comparative subdeletion, hidden degree, comparative clause ellipsis, more than someone has, than clause syntax often reveal confusion between three different phenomena: an unspoken degree in a comparative standard clause, a larger omitted predicate recoverable from context, and a comparative in which the two sides use different predicates or noun types. The practical task is not to settle every competing syntactic theory. It is to preserve the comparison’s truth conditions, its participants, its scales and any uncertainty in reconstructing the missing material.
This article is the specialist comparative deletion and subdeletion precision layer under eduKateSG’s Master Art of Translation. The English grammar background remains with How English Comparative Clauses Work, and broader translation of omitted material remains with Translate Ellipsis and Omitted Meaning. Here the owner is the degree relation inside comparisons, including what is pronounced, what is inferred, and when different measures are compared.
Quick Read
For “Maya ran faster than Theo did,” compare MAYA’S RUNNING SPEED with THEO’S RUNNING SPEED; “did” stands in for a relevant running predicate, and the second degree is not overtly repeated. For “Maya read more novels than Theo read articles,” compare COUNT OF NOVELS READ BY MAYA with COUNT OF ARTICLES READ BY THEO; the two noun kinds differ but the number dimension is shared. In every target, preserve the inequality or equality operator, scale, participant roles, temporal setting and the source’s stated or unstated scope.
One-sentence answer
Translate comparative deletion and subdeletion by reconstructing the implicit degree or quantity in the comparison clause, keeping the standard and scale fixed, and avoiding extra events or comparisons that the source never asserted. Restore enough hidden meaning to prevent ambiguity without mechanically copying English omissions.
Comparison is a relation between measured values
A comparative does not usually compare two people as undifferentiated wholes. “Rina is taller than Keon” compares Rina’s height with Keon’s height. “Rina works more carefully than Keon” compares manners or degrees of care associated with two performances. “Rina produced more sketches than Keon did” compares the number of sketches produced. The standard may be a person, state, object, event or expectation, but the comparison operates over a specified dimension.
The most useful editorial representation is VALUE 1, RELATION, VALUE 2, SCALE. For a taller-than sentence, the relation is strictly greater height. For “as tall as,” it is equality or a context-sensitive degree relation expressed through an equality construction. For “no taller than,” there is an upper-bound relationship rather than a simple greater-than claim.
A target language may lack an English-style than-clause. It could use a standard marker, comparative adjective, serial verb, exceed construction, or degree expression. The syntactic method is flexible; the ordered relation between values is not.
What comparative deletion means in the useful practical sense
Consider “This bridge is longer than that bridge is.” English does not ordinarily repeat “long” after the second “is” in the same way it would repeat an ordinary descriptive adjective in another construction. Analyses of comparative deletion treat a degree-bearing expression associated with the standard clause as unpronounced or otherwise represented by a silent dependency. The relevant meaning is that the first bridge’s length exceeds the second bridge’s length.
A literal reconstruction like “longer than that bridge is long” can be useful for private analysis, but it is not a natural public English translation. The translator’s notes may identify the missing scale while the released target sentence uses the target language’s usual comparative syntax. Importantly, some linguistic frameworks analyze comparative deletion through movement and deletion, while others differ in details; the translation invariant is the recoverable standard degree.
If “that bridge” is already established in context, copying its full noun phrase into the target may be unnecessary. If ambiguity arises because several bridges have just been mentioned, a fuller phrase may be needed. Economy and clarity must be balanced after the referents are secure.
Subdeletion: the second side can name something different
Now consider “Leila owns more novels than Sora owns biographies.” The quantity of novels owned by Leila is compared with the quantity of biographies owned by Sora. The two objects are different kinds, but both are counted. In analyses called comparative subdeletion, part of the measure expression in the standard clause is absent while lexical material such as the noun remains. Thus the target must not automatically replace biographies with novels for tidy parallelism.
A natural comparison can also relate dimensions of different properties when they share a coherent measurement scale: “The table is as long as the door is wide” compares two lengths, even though one is called length and the other width. The target must retain both the table’s length and the door’s width. Translating both as length changes the standard, not merely the words.
Because the technical label is theory-dependent, the reviewer should concentrate on the operational distinction: ordinary repeated-predicate reduction versus comparisons containing a different overt nominal or adjective in the standard. The latter demand explicit attention to which quantity the standard measures.
Deletion versus broader ellipsis
A sentence such as “Maya is taller than Kai” omits material beyond any degree dependency; there is no overt “is” in the second phrase. Comparative ellipsis can make an already degree-dependent clause shorter. “Maya is taller than Kai is” keeps the copula visible, whereas “Maya is taller than Kai” does not. Neither should be treated as a wholly different assertion merely because one contains an auxiliary.
With verbs, “Maya wrote more than Kai did” can omit a repeated verb phrase. But the comparison may concern number of words, amount of time, number of articles or some other quantity set by context. Restoring “Kai wrote” may not be sufficient to establish the measured dimension. First determine what “more” measures; then determine which predicate is recoverable.
A target may need to repeat the verb or noun to avoid a fragment. That expansion is legitimate if it restores supported content and does not introduce a new difference such as different subjects, objects or dates.
Scale identification before dictionary choice
English “more” can measure count, mass, time, intensity, frequency or degree. “More books” normally means a higher count of books. “More water” may mean greater quantity or volume, subject to context. “Worked more” is underspecified between hours, effort and output until discourse settles it. A translator who always maps “more” to one unmarked generic comparator risks making a target sentence incomplete or oddly broad.
Write the scale explicitly in the review worksheet: height, mass, amount, number, speed, price, error rate, probability, distance or duration. Then record the units if provided and note when the source leaves them unspecified. Do not infer a unit merely because a target phrase usually invites one.
In scientific prose, apparent synonymy can conceal different scales. Higher accuracy, greater precision and stronger recall may refer to different metrics. Maintain the exact metric and its direction: for some errors, a lower number is better performance, while for some scores a higher number is better.
The direction of comparison
“Team A finished sooner than Team B” says A’s completion time precedes B’s, given a shared reference. “Team A took longer than Team B” says A’s duration is greater. Translating both with a generic “better” or “more efficient” introduces evaluation that is not directly asserted. “Fewer errors” counts errors, whereas “less error” can be acceptable in some aggregate contexts but may change whether the metric is count or magnitude.
Comparative deletion makes this fragile because the full standard expression is not repeated to remind readers of the direction. Reconstruct it mentally: A finished earlier than B finished; A’s duration was longer than B’s duration. If a target language orders the standard before the comparee, verify that the inequality has not been reversed during reordering.
A simple back-translation is useful but insufficient on its own. Ask a bilingual reviewer to state, without looking at source word order, which value is larger, smaller, earlier, later or equivalent.
Participant roles in comparative clauses
“Mina likes poetry more than Arun” can mean Mina likes poetry more than Mina likes Arun, or Mina likes poetry more than Arun likes poetry, depending on context. The same visible words permit different elided clauses. Grammar and discourse may disambiguate, but the source might intentionally remain open.
When the stakes matter, list candidate reconstructions before translating. Do not choose the friendlier interpretation simply because a target sentence sounds more natural. If both readings remain genuinely available, a target that preserves the ambiguity is preferable where feasible; otherwise a translation note or editorial query may be necessary.
Technical literature has analogous problems: “The model predicted Group A better than Group B” might compare prediction accuracy for groups or suggest that one model was better than another predictor, depending on full context. Preserve the actual predication and object roles rather than smoothing away an unresolved comparison.
Comparative deletion with expectations and predictions
“The turnout was higher than organizers expected” compares actual turnout with an expected turnout level. It does not mean turnout was higher than organizers, nor that organizers performed the event. The hidden degree is a number of attendees anticipated before or during the event. A target that uses “above the organizers” metaphorically may be unintelligible or incorrect.
“The temperature fell further than the model predicted” involves a predicted extent of change. It may not mean the final temperature was lower than the predicted final value unless both share a reference baseline. Distinguish the magnitude of the fall from the endpoint itself. Translators should inspect variables, units and timepoints in surrounding charts.
Where a prediction is itself uncertain, do not add a fixed numerical threshold unless one is stated. “More than expected” tells us the observed value exceeded a relevant expectation, not what that expectation numerically was.
Cross-property comparisons and commensurability
“The desk is wider than the shelf is deep” compares desk width with shelf depth, two measures of length. This can be sensible because both are expressed on a compatible scale. The adjective labels differ, but they map into a common dimension. A mechanical system may normalize both adjectives to “wide,” losing which measurement belongs to the shelf.
Other pairs are not automatically comparable. Height versus salary, for example, does not form a meaningful single quantitative scale without a metaphor or explicit transformation. An unusual poetic comparison may deliberately exploit such mismatch; scientific writing normally should not. The translator should not silently invent a conversion or evaluation rubric to make an awkward original mathematically neat.
In professional reports, always preserve named dimensions even when the target requires a rephrased expression such as “The desk’s width exceeds the shelf’s depth.” This is a case where a nominalized target comparison can be clearer and more faithful than a direct adjective-to-adjective calque.
Comparative subdeletion with different object types
“Sadia sold more chairs than Ben sold tables” compares chair sales for Sadia with table sales for Ben. It does not say Sadia sold more chairs than Ben sold chairs. It also does not establish that the chairs and tables had the same prices or that revenue was greater. Count is the most natural scale unless the context establishes something else.
Keep both object categories and agents in a four-part record: seller A, category A, seller B, category B. When a target language requires classifiers, ensure that the classifiers suit each noun while a higher-level comparison still refers to quantity. If a translator selects one classifier for both categories merely because they appear in parallel positions, the target may become ungrammatical or alter the measure.
The same issue occurs with tests completed, documents submitted and devices inspected. A cross-category comparison can be relevant even when the categories are not identical. Do not flatten it into an apparently cleaner same-category comparison.
Count nouns, mass nouns and classifiers
“More bottles than cans” commonly compares counts of containers. “More water than juice” may compare quantities of liquids by volume, mass or servings depending on situation. In some target languages, count classifiers or measure words must appear explicitly. They should reflect the intended units, not convert an amount claim into a count claim without basis.
Suppose the source says “The tank contained more water than the backup vessel held.” The compared dimension may be volume or mass, but the target should not assert liters unless the document supplies that unit. Conversely, “The delivery included more water bottles than juice boxes” refers to packages, not fluid volume. Reading neighboring nouns and verbs matters.
Subdeletion adds another challenge when one side uses a count noun and the other a mass term. A valid comparison may require a measure phrase or contextual unit of equivalence. If the source does not define it, the translator should preserve the ambiguity or query the author rather than invent a precise amount.
Comparative standards across time
“She works faster now than she did last year” compares two states of the same person at different times. The subject is constant but the temporal anchor changes. A target that makes the standard a different employee destroys the trajectory. “The equipment vibrates less than before” may omit the earlier measurement event, but it still compares the present with a contextual baseline.
In historical analysis, “The city was larger than it had been before the reconstruction” compares sizes across periods. Do not assume identical administrative boundaries without evidence. The unit of comparison can shift in real-world descriptions, and a translator’s role is to preserve what the source claims, not repair the underlying research.
When translating tense and aspect, retain whether the earlier condition was habitual, ongoing or a completed measurement. “Had been,” “was,” and “used to be” can support different temporal interpretations. Do not erase those differences simply because the comparative adjective remains unchanged.
Equality comparisons and “as…as”
Comparative dependencies also occur in equality expressions. “The platform is as wide as the road is” compares two widths. “The table is as long as the partition is high” can compare two length measures named through different orientations. Unlike “longer than,” the central relation is equality or meeting a relevant degree standard, not strict superiority.
A frequent error is using the target’s greater-than form for an English “as…as” expression. Another is treating the first “as” as causal and the second “as” as temporal. Determine the construction as a unit, especially when nested adverbials make the sentence long.
There are pragmatic limits. “As good as” in certain idioms may convey nearly or effectively rather than literal exact equality. Technical translation should first check whether the source reports a measured equality, a qualitative judgment or an idiom. Each requires a different degree of literalness.
Less than, fewer than and reversed inequalities
“She used less fuel than he did” indicates a lower amount of fuel. “She made fewer attempts than he did” indicates a lower count of attempts. If a target language does not grammatically distinguish count and mass in the same way, the semantic dimension still must be preserved. Lower amount and lower number are related but not identical.
Negation changes the relation again. “No fewer than twenty arrived” may function as “at least twenty,” often with an implicature that the number is notable. “Fewer than twenty arrived” sets a strict upper limit below twenty. “Not more than twenty” sets an upper boundary at twenty, allowing equality. These are logical operators, not stylistic alternatives.
For regulatory or financial text, rewrite each operator into an inequality in the private worksheet and verify boundary inclusion. The comparative deletion question becomes secondary to exact logical scope, but the missing standard should still be identified correctly.
Modal verbs and hidden predicates
“Lina can read faster than Omar can” compares possible reading speeds under the stated ability frame. “Lina should have responded sooner than Omar did” involves an obligation or recommendation for Lina alongside a factual completed response by Omar. A mechanical expansion that gives both people the same modal can change what is asserted.
Distinguish event existence, capability, obligation, permission and prediction. Auxiliary substitution may conceal the underlying verb phrase, but it does not erase the modal’s meaning. If only one side has “might,” the target cannot legitimately spread that uncertainty over both sides.
A clinical trial report may say “The treatment might reduce symptoms more than the placebo does.” Here potential treatment effect is compared with a reported or assumed placebo effect. Preserve both the uncertain treatment claim and the chosen comparison standard; do not turn them into a demonstrated superiority trial unless the source supports that conclusion.
Negation and the scope of comparison
“She is not taller than her sister” ordinarily denies the strict taller-than relation; it does not necessarily assert that she is shorter. Equal height is compatible. Similarly, “The new device is not more reliable than the old one” does not prove the new device is less reliable. A translation that swaps “not more” with “less” strengthens the claim.
Comparatives can also contain negation within the standard: “The device performed better than we had not expected” is awkward and needs careful interpretation; real texts may involve negative expectations or conditional assumptions. Do not fix odd source grammar by guessing the author’s intent without marking the uncertainty.
For “No more students passed than failed,” parsing the quantifiers and comparison is crucial. Is it a statement that the number passing did not exceed the number failing, or a rhetorical claim in a broader context? Record the two sets and the inequality explicitly. Natural language often expresses thresholds less transparently than symbolic form.
Subdeletion versus coordination
“Maya bought more pencils and Omar pens” may be a coordination with an omitted verb in the second conjunct, not a than-clause comparison. “Maya bought more pencils than Omar bought pens” has an explicit comparison marker and a degree relation. Similar noun pairs can mask quite different constructions.
Do not apply comparative deletion analysis to every sentence with unequal quantities. A contrastive conjunction might list actions rather than rank them. Conversely, a comparative standard clause can contain coordination: “More students attended than registered and paid.” The translator then needs to determine whether one or two predicates are included in the comparison.
A clear syntax map prevents wrong operator placement. Identify the comparative marker and its scope before reconstructing individual omitted predicates. If the clause structure is genuinely ambiguous, preserve the relevant readings or seek clarification before release.
Comparative standards inside subordinate clauses
“More applicants qualified than the reviewer believed would qualify” includes an attitude holder and a projected number. The comparative standard belongs to the reviewer’s belief, not to the reviewer as a person. The clause may contain further embedded material that carries its own tense, modality or negation.
Nested comparisons can become especially difficult: “The result was closer to the prediction than the team had thought it would be.” Here closeness, prediction, and prior expectation interact. The translator must preserve whether the team predicted the numerical result or merely estimated how closely the result would match another prediction.
When needed, draw a three-level diagram in the working notes: observed value, comparison property, mental-state or prediction standard. The public target need not show that diagram; it needs to remain interpretable without it.
Units, rounding and mathematical comparators
A sentence such as “The sample weighed more than the control did” compares weights, but units may be implied in the scientific context. A target that supplies kilograms when the table shows grams risks enormous error. Unit conversion should occur only under a defined publishing convention and should be checked numerically.
“About ten percent more” differs from “ten percentage points more.” The first can express a relative change, the second an absolute difference between percentages. Comparative syntax does not erase the distinction. Similarly, “twice as many” and “two more” are not interchangeable. When the comparison uses a hidden standard, reconstruct the baseline before touching the arithmetic.
Rounding matters at boundaries. “Greater than 0.5” and “at least 0.5” treat equality differently. An apparently small translation change in threshold language can alter eligibility or measurement outcomes. Use an independent numeric QA check for high-stakes comparisons.
Worked case: science equipment dimensions
Source: “The sample chamber is deeper than the inlet is wide.” A mechanical editor might treat both adjectives as describing depth, especially if the target prefers one comparative construction with repeated adjectives. But the comparison uses chamber depth versus inlet width. The shared measurement scale is length. A reliable paraphrase is “The depth of the sample chamber exceeds the width of the inlet.”
This translation is clear if the context is technical and direct comparison across dimensions is intended. It must not imply that chamber depth exceeds inlet depth, or that the inlet is narrower than a different inlet. Labels and dimension orientation should match diagrams. In a drawing revision, verify whether both features are measured in the same units before converting any numerals.
Worked case: library inventory
Source: “The school borrowed more atlases than the college borrowed dictionaries.” The comparison concerns the number of atlases borrowed by the school versus dictionaries borrowed by the college. Nothing in the sentence tells us which institution borrowed more items overall. A translation that simplifies the message to “The school borrowed more books than the college” overgeneralizes both categories.
A classification system may call atlases reference books and dictionaries reference works, but those taxonomies are not permission to broaden the claim. Preserve each named object type, the two institutions and the borrowed count. If the original quantifies over copies rather than distinct titles, do not substitute titles without evidence.
Worked case: exam marks and time
Source: “Asha solved more problems than Ravi completed in the time allowed.” The comparison could be understood as counts of problems solved and completed, but the wording may leave whether “solved” and “completed” represent the same achievement ambiguous. A translator should not silently replace one verb with the other to enforce parallelism.
If the exam uses attempted, completed and correct as separate metrics, their distinction is crucial. Reconstruct the count for each predicate, keep the timed setting attached to the intended side, and query the author if the comparison is not mathematically well defined. Information that sounds symmetrical in everyday prose can encode different assessment categories.
Worked case: speed versus duration
Source: “The courier drove faster than the dispatcher expected.” The scale concerns speed relative to an expectation, not delivery duration unless the context links them. A version such as “The courier arrived earlier than predicted” is plausible but not entailed: a faster speed does not guarantee an earlier arrival if departure time or route length differs.
Write observed speed and predicted speed as the two compared quantities. If the source only says “drove faster,” preserve that activity. The translator may restructure an implicit expectation as “at a greater speed than the dispatcher had anticipated,” but should not invent a timetable or a delivery target.
Worked case: the empty control group
Source: “The experimental group had more positive responses than the control group did.” The reader compares counts or rates depending on the study’s operational definition. If group sizes differ, a raw number comparison is not necessarily a rate comparison. A translation that changes “positive responses” into “higher response rate” can distort the evidence.
The target should preserve the unit that the source actually states. If the study later reports normalized percentages, use those where appropriate but do not retroactively change the meaning of this sentence. Distinguish controls as people, samples or measurement conditions, depending on context.
Worked case: project cost
Source: “The repair cost more than the contractor estimated.” This compares actual repair cost with a prior estimate. The contractor is an estimator, not the thing measured or a rival seller. The translation must preserve the fact that the standard is estimated cost rather than the contractor’s own income.
If currency appears elsewhere, do not infer a new one from locality. If the contract has taxes and exclusions, do not silently decide whether the comparison is tax-inclusive. A clear target may explicitly say “The repair’s actual cost exceeded the contractor’s estimate,” retaining the hidden comparison and its source.
Worked case: a cross-category sales report
Source: “Mina shipped more monitors than Hari shipped printers.” The two noun kinds are different but the likely shared scale is a number of units shipped. A literal target that accidentally creates “more monitors than printers shipped by Hari” can obscure the first agent. A short table in translation notes resolves it: Mina—monitors—shipped; Hari—printers—shipped.
Unless the report says otherwise, do not infer equal average unit prices or compare revenues. Shipping count, sales count, orders and inventory balance are different quantities. Subdeletion is not license to normalize everything into “more sales.”
Worked case: matching dimensions
Source: “The panel is as high as the doorway is wide.” Two different physical orientations are mapped onto compatible linear measurements. A target translation using “the panel is as wide as the doorway” would be fluent but incorrect. The panel’s height and the doorway’s width are the intended variables.
Technical diagrams should confirm which faces and axes are measured. If the panel is to be rotated when installed, that engineering inference belongs in a separate instruction, not in a translation of the comparison. The equality should remain just the statement the source makes.
Worked case: a mistaken numerical boundary
Source: “The replacement battery should cost no more than forty dollars.” This is an upper price bound allowing forty dollars. A version saying “less than forty dollars” excludes exactly forty and changes purchasing eligibility. The presence or absence of a reduced than-clause is less important here than the logical inequality.
When localizing currencies, never treat the number as a conversion task unless requested and a rate/date policy is defined. A literal currency amount is not automatically transferable to another market’s purchasing conditions. Keep the operator and the monetary unit stable unless the approved localization brief directs otherwise.
Worked case: comparative ambiguity in a personal sentence
Source: “Nora admires Sam more than Talia.” One interpretation compares Nora’s feelings for Sam and Talia. Another compares Nora’s admiration for Sam with Talia’s admiration for Sam. The surface phrase permits both in the absence of clarifying context.
A target language may require an explicit verb, case marker or pronoun that resolves the ambiguity. That is a translation problem, not a reason to guess. Read surrounding dialogue for a licensed reading. If neither is established and ambiguity matters, record both analyses for an editor. The goal is not always to remove ambiguity; it is to avoid choosing a meaning the author did not commit to.
Worked case: before-and-after comparison
Source: “The student wrote more accurately than before.” The comparator is the student’s earlier writing accuracy, not that of another student or a teacher. A target that turns “before” into a date should retain uncertainty if no timepoint is supplied.
Accuracy is not necessarily speed or vocabulary sophistication. In an educational report, a faithful translation may expand the short sentence to “The student’s current work contains fewer accuracy errors than their earlier work” only if the metric really is error count. If the source uses a holistic assessment score, that expansion may be wrong. Preserve the named performance dimension.
Worked case: counterfactual baseline
Source: “The team would have finished sooner than it did if the files had arrived.” The result is counterfactual, conditional on files arriving; the observed completion time is the comparison standard. A translation that states “The team finished sooner” removes the unreality and reverses the actual timeline.
Map ACTUAL FINISH versus HYPOTHETICAL FINISH and mark the unrealized condition. The hidden degree in the comparative clause is the actual finish time, while the matrix projects an earlier counterfactual time. All three components—the condition, comparison and actuality—must survive. This is why a purely lexical more-than translation is insufficient.
A disciplined translation workflow
Begin by finding the comparison marker and scale. Identify the compared item or event, the standard item or event, and the dimension over which the relation operates. Next reconstruct any omitted predicate and underlying degree in notes. If the standard contains a different noun or adjective, preserve that difference explicitly. Check for modal, negation, tense, quantifier and boundary operators.
Draft a target sentence in a native comparative construction. Then test it with a paraphrase naming the two complete measured values. If the source compares a bridge’s length to a road’s width, does the target make those two measurements clear? If the source compares expected and actual attendance, does the target retain the expectation holder and the actual event? If both conditions pass, simplify as far as naturalness allows.
For numerical or scientific material, add unit and inequality checks. A second reviewer should independently name the direction of comparison and verify that no category, participant or scale has changed.
Frequent failures and repairs
The most common failure is false same-category parallelism: “more novels than biographies” becomes “more novels than novels.” Restore the overt noun in the standard. Another is comparison reversal after word-order changes; write the inequality explicitly and verify which side exceeds the other. A third is adding an absolute number when the source only supplies a relative standard. Remove guessed numerals.
An ambiguity failure occurs when a target language requires a hidden verb and the translator chooses one without contextual evidence. Review alternative reconstructions. A metric failure occurs when count becomes rate, accuracy becomes precision, or width becomes depth. Recheck the scale and units. Finally, a modal failure turns a predicted or hypothetical comparison into a factual result. Restore the timeline and level of certainty before editing style.
Practice: reconstruct the silent measures
Analyze five sentences before attempting any target language. “The tower is higher than the fence is” compares two heights. “The door is narrower than the desk is deep” compares door width and desk depth. “The team filed more claims than the rival team filed appeals” compares counts of different filings. “The observed rate was lower than predicted” compares a rate with its predicted value. “The adjustment did not reduce errors more than the prior adjustment did” denies a greater error-reduction relationship without necessarily saying the new adjustment increased errors.
For each, label the two values, the operator, any hidden verb and any assumption the source does not establish. Translate into a natural target comparative and back-paraphrase. The exercise is successful when both the precise measurement relation and the original uncertainty are preserved.
Frequently asked questions
What is comparative deletion?
In influential linguistic analyses, it is a dependency in comparative standard clauses where a degree expression is unpronounced. For translators, the practical consequence is a hidden measure that must still be understood.
What is comparative subdeletion?
It is a related construction in which different lexical material can remain in the standard, as in comparisons of books with magazines or the length of one object with the width of another.
Is comparative deletion the same as omitting a repeated verb?
Not exactly. Broader comparative ellipsis can remove a repeated predicate or auxiliary material in addition to an underlying degree dependency.
Must a translation restore all missing words?
No. It must restore meaning. A natural target comparative may omit material differently or use a nonclausal standard.
How do I identify the correct scale?
Ask whether the sentence compares height, count, amount, speed, duration, price, probability or another measurable property. Context may be required.
Why can “more than” be ambiguous?
The comparison standard or predicate can be omitted. More than one plausible reconstruction may fit the visible wording.
Does “not more than” mean “less than”?
No. The former commonly allows equality at the upper boundary; the latter is strictly below it.
Can a comparison relate different adjective dimensions?
Yes when they map onto a meaningful common measure, such as one object’s height and another object’s width, both measured as lengths.
Position inside the eduKate translation architecture
The Master Art of Translation governs meaning preservation. The English owner How English Comparative Clauses Work explains the grammar, while Translate Ellipsis and Omitted Meaning explains broad reconstruction discipline. This article owns the distinct comparative degree and standard relation, particularly where the standard names a different noun or adjective.
For specialist study, see Comparative Deletion and Comparative Subdeletion by Winfried Lechner and Norbert Corver, and the descriptive examples in Taalportaal’s comparative subdeletion entry. Linguistic mechanisms remain debated; translation verification should foreground the observable semantic comparison.
Final principle
The visible words around “than” or “as” do not always spell out the measure being compared. A translator protects the hidden degree by making both sides of the relation explicit in analysis, preserving the exact scale and standard, and then writing natural target-language comparative grammar. Concision is safe only when the same comparison remains recoverable.