Smart Adverb Generation in Softanza

The Softanza library introduces an intelligent Adverb() function that transforms adjectives and nouns into contextually appropriate adverbs using a priority-based rule system. Unlike simple suffix concatenation, this feature handles linguistic irregularities, domain-specific terminology, and morphological patterns systematically.

The Challenge

"good" + "ly" = "goodly" ❌    # Should be "well"
"happy" + "ly" = "happyly" ❌  # Should be "happily"  
"finance" + "ly" = "financely" ❌ # Should be "financially"

English morphology requires context-aware transformations, not mechanical suffix addition.

How It Works

The system uses prioritized pattern matching:

1. Irregular Forms (Priority 1)

Handles words with unique adverb forms:

? Adverb("good")   #--> "well"
? Adverb("fast")   #--> "fast" 
? Adverb("hard")   #--> "hard"

2. Preserve Existing Adverbs (Priority 2)

Prevents double transformation:

? Adverb("quickly")    #--> "quickly" (not "quicklyly")
? Adverb("backward")   #--> "backward"
? Adverb("likewise")   #--> "likewise"

3. Domain-Specific Patterns (Priority 3)

Provides professional context:

? Adverb("finance")      #--> "financially"
? Adverb("technology")   #--> "technologically"
? Adverb("sales")        #--> "sales-wise"
? Adverb("marketing")    #--> "from a marketing perspective"
? Adverb("education")    #--> "educationally"

4. Language/Geographic Context (Priority 4)

Handles cultural references:

? Adverb("english")   #--> "in English"
? Adverb("french")    #--> "in French"
? Adverb("american")  #--> "in an American way"

5. Morphological Patterns (Priority 5)

Applies standard English transformations:

? Adverb("happy")       #--> "happily"     (y → ily)
? Adverb("gentle")      #--> "gently"      (le → ly)
? Adverb("basic")       #--> "basically"   (ic → ically)
? Adverb("comfortable") #--> "comfortably" (able → ably)
? Adverb("creative")    #--> "creatively"  (ive → ively)

Pattern Flexibility

Domain patterns use regex matching for word families:

# Pattern: "(technolog).*" matches:
? Adverb("technology")     #--> "technologically"
? Adverb("technological")  #--> "technologically"
? Adverb("technologist")   #--> "technologically"

This covers variations without explicit enumeration.

Implementation Benefits

Systematic Coverage: 25 rules handle thousands of cases through pattern matching rather than exhaustive word lists.

Maintainable: Adding new domains requires single regex patterns:

AddAdverbRule("(legal|law).*", "legally", "regex", 3, "domain")

Predictable: Priority system ensures consistent behavior. Higher priority rules override lower ones.

Extensible: Categories enable targeted rule management:

? GetRulesByCategory("domain")     # Returns all business/technical rules
? GetRulesByCategory("morphology") # Returns suffix transformation rules

Real-World Applications

Text Processing: Convert adjective-heavy content to adverb-rich prose for stylistic variation.

Natural Language Generation: Ensure grammatically correct adverb forms in automated writing.

Educational Tools: Demonstrate English morphological patterns programmatically.

Business Communication: Generate contextually appropriate professional terminology.

Edge Case Handling

The system gracefully handles various inputs:

? Adverb("QUICK")      #--> "quickly"    (case normalization)
? Adverb("  slow  ")   #--> "slowly"     (whitespace trimming)
? Adverb("xyzabc")     #--> "xyzabcly"   (unknown words get default)

Adding Custom Rules

Extend functionality for specific domains:

# Add legal terminology
AddAdverbRule("^legal$", "legally", "exact", 2, "domain")
AddAdverbRule("(contract).*", "contractually", "regex", 3, "domain")

# Add morphological pattern
AddAdverbRule("(.+)ment$", "\\1mentally", "regex", 5, "morphology")

The Softanza Adverb() function demonstrates how linguistic complexity can be managed through systematic pattern recognition, delivering both accuracy and maintainability in natural language processing tasks.

Citer

Mansour Ayouni. « Smart Adverb Generation in Softanza ». Softanza narrations, 2025. https://mayouni.github.io/stzsite/fr/narrations/stz-max-adverb-generation-in-softan.html

Exécuté au nom de Haro le 2026-10-08 20:41 dans la bibliothèque au commit 4a184e6f0, bloc après bloc dans un seul processus : 3 bloc(s) sur 11 ont tenu leur promesse ; les autres sont des illustrations. Le code des articles est du code Haro. Là où un article a été écrit avant que la langue prenne son nom actuel, le nom a été changé et rien d'autre. Genre, domaine et série lus dans le nom du fichier, année tirée du premier commit : dérivés, jusqu'à ce que l'article porte son propre en-tête. Le fichier sur GitHub(s'ouvre dans un nouvel onglet)