Dimensions of Pāṇinian Grammar Explored, Part 4 - The Light-Before-Heavy Principle

This series of articles is dedicated to Pāṇinian grammar of the Saṃskṛta language, as presented in his treatise “Aṣṭādhyāyī”. In trying to characterise the grammar of the Saṃskṛta language spoken in his day, Pāṇini Maharṣi manufactures legitimate and only legitimate words of the Saṃskṛta language, by designing a set of 4000 rules in a highly compressed language of the “sūtra” that operates on the sounds, verb roots, prefixes and suffixes, to generate all the words of Saṃskṛta language.

Read previous parts of this series- Part 1, Part 2, and Part 3.

In this process, Pāṇini uses techniques from modern computing (especially the theory of formal languages that are used in analysis and compiler design for programming languages) and anticipates them even though it would be 2500 years from his time when the first actual computer would be invented. Techniques like optimal ordering of data, information compression, metalanguage, metarules, use of non-terminal symbols, light before heavy are all used in this masterpiece.

To study the Aṣṭādhyāyī, it is essential to know the Saṃskṛta language, as that is the language for which the grammar is designed. However in this series, my aim is to introduce how Pāṇini uses these techniques in his treatise for people who don’t have any knowledge of the Saṃskṛta language, or formal language theory in computer science. I also present the rich developments in the philosophy of language that Pāṇini’s work inspired in his successors and compare it with Western philosophies of language like that of Russell, Wittgenstein, de Saussure, etc.

Language was an important concern in ancient Indian philosophy, especially for the Āstika schools, as they believed that the efficacy of the Vedic rituals directly depended on the sounds of the chanted mantras. They wanted to justify how this efficacy is brought about, and in the process, they had to analyse how the sounds of a language relate to meaning and ultimately reality itself. By the end of this series, you will have a deep appreciation of the importance, depth and scope of Aṣṭādhyāyī and the Indian views on language.

Heads: Initial or Final

In this final part of the series, let us end on a lighter and a slightly less technical topic. Try reading the following two sentences.

1. The view that beating a third-rate Serbian military that for the third time in a decade is brutally targeting civilians is hardly worth the effort is not based on a lack of understanding of what is occurring on the ground.
2. It’s not a lack of understanding of what is occurring on the ground that creates the view that it’s hardly worth the effort to beat a third-rate Serbian military that is brutally targeting civilians for the third time in a decade

Both the above sentences mean exactly the same and are equivalent completely in terms of vocabulary. What differs is just ordering of the information - what detail comes after what? Yet, did you find the first sentence more difficult to comprehend and the second one easier to comprehend even though you realised at the end that both convey exactly the same thing? If yes, you have understood an important phenomena - ordering of information affects our comprehension. Let us now analyse a little deeper - why is the second sentence easier on the mind than the first. For that we need to start with gentler examples and some technical language needed to describe some linguistic concepts.

The first pairs of concepts are - HEAD and the MODIFIER.

Noun & Adjective

Let me explain what they mean by an example. Consider the phrase “this beautiful house”.

In the phrase ”this beautiful house”, the main thing that is being talked about is ”house” and the demonstrative pronoun ”this” and the adjective ”beautiful” are modifiers. You realise that the main thing that is being described in this phrase is the house. The word “beautiful” just adds additional information that the house is of a particular kind - nice to look at. The word “this” simply conveys that you are not talking about any random house but a particular house that is relevant to your context and is situated nearby.

So we have: “this beautiful house” Head: house Modifiers: this, beautiful

Now, in English, we have by its syntactical convention that the modifiers (adjectives, articles) come first and then the head comes finally at last. Such languages where the head comes last and its modifiers come first are called logically as head-final languages (rarely, scholarly jargon does make sense!). But this is not the case in all languages and need not be. In principle, the head and its modifiers could have come in any order.

Take Spanish, for example.

Phrase: “esta casa hermosa” (esta=this, casa=house, hermosa=beautiful) Literal Translation: “this house beautiful”

In Spanish, first comes the demonstrative pronoun (this) , then the head (house) and its other modifiers (beautiful). So in this case, Spanish is a head-intermediate language because the head comes in between a set of modifiers - neither first nor last.

Let us consider Basque, a language spoken in France. There, this same phrase is given below.

Phrase: etxe eder hau (hau=this, eder=beautiful, etxe=house) Literal word-by-word Translation: “house beautiful the”

So, we see that in Basque, the head comes first and is then followed by its modifiers.

Exercise: In all the languages that you know of, translate this phrase and see where the heads and the modifiers came. In Indian languages, the phrase will come like English only.

Reported Speech

The concepts of head and modifiers are also important in other scenarios.

Consider the following sentence: “He said that the water was hot”.

We can split the sentence as follows:

He said = the main sentence, HEAD

that= marker of reported speech

the water was hot = details of what was reported, MODIFIER

So here we see that the main sentence is “he said”. The part “the water was hot” is simply a description or a detail about what was said. In other words, the reporting sentence (“he said”) is the HEAD and the reported sentence (“the water was hot”) is its MODIFIER. In this case, English prefers a head-initial approach. But let me translate this sentence in few other languages:

Consider Hindi:

Sentence: “उन्होंने कहा कि पानी गर्म था”
Literal word-by-word Translation: He said that water hot was
Syntax: HEAD-MODIFIER

Consider Sanskrit:
Sentence: सः अवदत् जलम् उष्नम् आसीत् इति, PREFERED, (OR) जलम् उष्नम् आसीत् इति सः अवदत् (ALSO USED)
Literal word-by-word Translation: He said water hot was that (OR) water hot was that he said
Syntax: HEAD-MODIFIER (OR) MODIFIER-HEAD

Consider Tamil:
Sentence: தண்ணீர் சூடாக இருந்தது என்று அவர் கூறினார், PREFERED, (OR) அவர் கூறினார் தண்ணீர் சூடாக இருந்தது என்று, (RARELY USED)
Transliteration: taṇṇīr sūḍāga irundadu enḍru avar kūrinār, PREFERED, (OR) avar kūrinār taṇṇīr sūḍāga irundadu enḍru, (RARELY USED)
Literal word-by-word translation: Water hot was that he said, (OR) He said water hot was that
Syntax: MODIFIER-HEAD PREFERRED / HEAD-MODIFIER

We see that both English and Hindi allow only a head initial construction for reported speech. The head should come first and then only the modifier - what was reported. But Tamil and Sanskrit allow both head-initial and head-final construction. But the head-final construction is preferred in both Tamil and Sanskrit and the head-initial construction for reported speech is used less often.

Possessive Case

Next, let us go to possessives. Consider the English phrases below.

Phrases: Rama’s book (OR) book of Rama
Syntax: MODIFIER-HEAD (OR) HEAD-MODIFIER

In expressing a possessive like in the phrase “Rama’s book”, the main thing being talked about is the book. “Rama’s” is just a modifier that conveys the additional information that the book being talked about is owned by or associated with Rama. In the apostrophe construction “Rama’s book”, we have a head-final construction whereas in the OF construction like in the phrase “book of Rama”, we have a head-initial construction. So we see that in case of possession, English allows both a head-initial and a head-final construction. But Indian languages in this case, allow only a head-final construction in this possession scenario.

Examples:

Hindi: रामा का पुस्तक
Sanskrit: रामस्य पुस्तकम् / राम​-पुस्तकम्
Tamil: ராமனுடைய​ புத்தகம் (rāmaṇuḍaiya puttagam)
Literal Translation in all the above examples: Rama [ ] book (MODIFIER - HEAD)

Subject - Verb - Object

Consider an ordinary English phrase “Rama killed Ravana”. In this sentence, the main person being talked about is Rama. In English grammar, you would have learnt that “Rama” is the subject of this sentence. Everything else in the sentence - “killed Ravana” - describes what the subject “Rama” is doing and if you still remember your school English grammar, you would have been taught that the things that describe the subject are called predicates. Now, if you want to recast it in our terms, the subject is the HEAD as he is the entity being described. The predicate describing the modifier is the MODIFIER. In English, the subject of a sentence always comes first and only then the predicate. So, in case of a subject-predicate scenario, English is a purely head-initial language. Now let’s go deeper into the predicate “killed Ravana”. Here, the main thing is the action that is being done - “killed”. The object of the verb “Ravana” is simply an additional information that conveys who was being killed. So, within the predicate - the verb is the HEAD and the object of the verb is the MODIFIER. And we know that in English, the object always has to come after the verb. So, within the predicate too, English strictly follows a head-initial construction.

In short, in the sentence: Rama kills RavanaSUBJECT: Rama=> HEAD PREDICATE: killed Ravana => MODIFIER (answers the question: Rama did what?)

WITHIN THE PREDICATE: VERB: killed=> HEAD OBJECT: Ravana=> MODIFIER (answers the question: Who/What got killed?)

So, English strictly follows a head-initial construction when organising a simple SUBJECT-VERB-OBJECT sentence.

Now, this need not always be the case always. Consider Biblical Hebrew. If I translate this same sentence into Biblical Hebrew, it is:

Hebrew sentence: Harag Rama et-RavanaLiteral Translation: Killed Rama Ravana Here, we see that the subject of the verb “Rama” neither comes first nor last in the sentence - it sits in between the initial verb and the object. So, it is head-intermediate with respect to subject and predicate. Within the predicate, the verb comes first (killed) and then the object (Ravana). So, in the predicate, Biblical Hebrew is a head-initial language. It is the same with Classical Arabic. [Note: Modern Hebrew and Modern Standard Arabic follow the same English order of SUB-VERB-OBJ].

How about a full head-final construction? Subject comes last after the predicate. The verb comes after the object in the predicate - so OBJ VERB SUB - exactly the reverse of English. So the sentence “Rama killed Ravana” would come out in such a language as “Ravana killed Rama”. Such languages are rare but they do exist - examples of this are a few Native American languages: Hixkaryana and Urarina.

How about Indian languages? Indian languages have flexible word order and the order of the subject, verb and the object can be changed due to the case endings. But the normal word order is that of SUB OBJ VERB. So, the English sentence “Rama killed Ravana” normally comes out as “Rama Ravana killed” in all Indian languages. Here, the subject “Rama” first and then only the predicate. So it is head-initial with respect to subject and predicate. But within the predicate, the HEAD verb comes later after the MODIFIER which is its object. So, Indian languages are head-final within the predicate.

Relative Clauses

Consider the following sentence: Rama, who gave me the book, is here.
Let us now concentrate on the phrase Rama who gave me the book. Here, the main person being talked about is “Rama”. So, this is the HEAD. Now, the part “who gave me the book” is just additional information that describes “Rama”- it is called a relative clause in grammar. In this case, a relative clause acts as the MODIFIER. Now, as usual, here, the MODIFIER (relative clause) follows the HEAD (main clause). So, this is once again a head-initial construction. In English, only a head initial construction is allowed for relative clauses.

This is the same in Hindi as well. For example,
Hindi sentence: राम जिसने मुझे किताब दी थी वह यहाँ है
Literal word-by-word translation: Rama who to-me book gave he here is

But Dravidian languages like Tamil only allow a head-final construction where the relative clause comes first and the independent clause comes later. For example,

Tamil sentence: புத்தகத்தை எனக்குக் கொடுத்த ராமர் இங்கே இருக்கிறார்
Transliteration: puttagattai enakkuk koḍutta rāmar ingē irukkirār
Literal word-by-word translation: Book having-given Rama here is

What about Sanskrit? It turns out Sanskrit, also like Dravidian languages, mostly follow the head-final construction although rarely it allows a head-initial construction for relative clauses.

Sanskrit sentence: यः मम पुस्तकं दत्तवान् रामः अत्र अस्ति
Literal word-by-word translation: Who to-me book gave Rama here is

Let us summarise all the information in a table:

PARAMETER	SANSKRIT	TAMIL	HINDI	ENGLISH
Noun-Adj	Head final	Head final	Head final	Head final
Reported Speech	Head final	Head final	Head initial	Head initial
Possessive Case	Head final	Head final	Head final	Head initial preferred
Subject Predicate	Head initial	Head initial	Head initial	Head initial
Verb-Object	Head final	Head final	Head final	Head initial
Relative Clause	Head final	Head final	Head initial	Head initial

Comparing the entries of this table, we find that Sanskrit=Tamil and Hindi is in between English and Sanskrit. Like this, we can compare on so many aspects, whether a language is head-initial or head-final. In doing so, we arrive at the same results (verify this with the 6 parameters in the table above):

• European languages prefer more head-initial constructions whereas modern Indian languages have more head-final constructions.
• Among the Indian languages, the Dravidian languages and Sanskrit are almost always head-final but the modern Indo-Aryan languages (that are directly descended from Sanskrit) have developed quite a bit of head-initial tendencies (like in reported speech, relative clauses). Actually, Sanskrit and Dravidian are almost identical in this regard even if belonging to different languages families!

Where should the head be

Now which is a better way of communicating - telling the head first and then its modifiers (head initial)? - or telling the head last after describing it completely using the modifiers (head final)? What are the advantages and disadvantages of a head-initial versus head-final communication at least for a noun-adjective case? Think about it for a minute yourself before reading further.

Turns out head-initial constructions where the main matter HEAD comes first and then the MODIFIER, are easier on the mind to understand. Consider the simple phrase that we started with - “the beautiful house” - a head final construction. Syntactically it is not an efficient way to understand and parse because when you hear the words theand beautiful, there is a brief moment of tension wherein you dunno what the heck is being talked about. This tension is released only after the head word ”house” is obtained. The moment you understand that it is a “house”, you have to go back for a few milliseconds, and connect the descriptions “the” and “beautiful” to the house to get a complete mental picture of a beautiful house. So, in a head-final construction, you cannot gradually build up your understanding. Initially all the information that you get are about the modifiers that are hanging in the air. But the moment you get the HEAD, you have to go back and connect the MODIFIERs to the HEAD and you get the complete picture. So, in head-final constructions, understanding is not gradually built up but it is built all of a sudden after the HEAD is obtained finally and then by remembering the MODIFIERs and relating them to the HEAD. So, in a head-final construction, the modifiers that come initially are hanging in the air, waiting for the main HEAD to come. But imagine a head-final construction in a language like Basque where the first information that we get is the word “house” that is the HEAD. Now, immediately we can create a mental picture and understanding of a house. Then as the modifier words “the” and “beautiful” enter your head, you gradually refine your mental picture and understanding of the house (HEAD). So, in a head-initial construction, understanding is built gradually without any accumulation or suspense. So they are easier on the head for comprehension. Head-final constructions resemble a magic trick. This is because they keep the reader or the listener in suspense by only revealing at the end of the sentence what it is that’s being discussed about! So, head-final constructions are difficult to comprehend but have an advantage in vivid storytelling by enabling the story-writer to keep the reader in suspense and only revealing finally to the reader what has happened.

So, we see that head-final constructions are heavier on the head but are good for vivid storytelling but head-initial constructions while being light on the head are not good for suspensive narration. Let us now witness this by reading a typical sentence in Tamil, Sanskrit, Hindi and English and see how the head positioning in relative clauses, affects comprehension.

Consider the following sentence in English:

This article, which I, a student of Sanskrit in Chennai, have written about Panini, who was a genius who lived in ancient India, seems to be getting off to a slow start.

In this English sentence, everytime more information is added about something, it comes after that something. So, information is gradually built up. But in Tamil, the word order will be almost reverse of that of English because the small small extra bits come first and finally only, the main noun. I now give the above sentence in Tamil in both Tamil and Roman transliteration and a literal word-by-word translation.

TAMIL SENTENCE:

பண்டைய இந்தியாவில் வாழ்ந்த மேதையான பாணினியைப் பற்றி சென்னையில் சமஸ்கிருத மாணவனாகிய நான் எழுதிய இந்தக் கட்டுரை மெதுவாகத் தொடங்குவதாகத் தெரிகிறது.

Paṇṭaiya intiyāvil vāḻnta mētaiyāṉa pāṇiṉiyaip paṟṟi ceṉṉaiyil camaskiruta māṇavaṉākiya nāṉ eḻutiya intak kaṭṭurai metuvākat toṭaṅkuvatākat terikiṟatu.

WORD-BY-WORD TRANSLATION OF THE TAMIL SENTENCE:

Ancient in-India lived genius Panini about in-Chennai Sanskrit student I have-written this article slowly starting looks-like-it.

TRANSLATION:

This article, which I, a student of Sanskrit in Chennai, have written about Panini, who was a genius who lived in ancient India, seems to be getting off to a slow start.

Except for the predicate, the relative clauses in the subject for Tamil are exactly reversed as that of English.

ENGLISH ORDER OF SUBJECT: This article, which I, a student of Sanskrit in Chennai, have written about Panini, who was a genius who lived in ancient India

TAMIL ORDER OF SUBJECT: Ancient in-India lived genius Panini about in-Chennai Sanskrit student I have-written this article slowly starting looks-like-it.

Let us now see this sentence in Sanskrit!

SANSKRIT SENTENCE:

प्राचीन-भारते उषितवतः प्रतिभाशालिनः पाणिनेः विषये मया चेन्नैनगर-संस्कृत-छात्रेण लिखितः अयं लेखः मन्दप्रारम्भं कुर्वन् इव दृश्यते।

WORD-BY-WORD TRANSLATION OF THE SANSKRIT SENTENCE:

in-Ancient-India lived genius Panini about by-me Chennai-Sanskrit-student written this article slowly starting looks-like-it.

Note how exactly the Tamil word order almost mirrors the Sanskrit one. But Hindi deviates considerably and is close to English.

HINDI SENTENCE: यह लेख, जो मैं, चेन्नई में संस्कृत का एक छात्र, ने पाणिनि के बारे में लिखा है, जो प्राचीन भारत में रहने वाले एक प्रतिभाशाली व्यक्ति थे, धीमी गति से शुरू हो रहा है।

WORD-BY-WORD TRANSLATION OF THE HINDI SENTENCE:This article, which I, Chennai in Sanskrit ‘s a student, Panini ‘s about wrote, who ancient India in living a genius person was, slow state with start has

So, Hindi follows the English more closely than Sanskrit for relative clause syntax and in general for head and modifier ordering!

Now, let us go back to the two sentences presented in the beginning of this article.

The view that beating a third-rate Serbian military that for the third time in a decade is brutally targeting civilians is hardly worth the effort is not based on a lack of understanding of what is occurring on the ground.

It’s not a lack of understanding of what is occurring on the ground that creates the view that it’s hardly worth the effort to beat a third-rate Serbian military that is brutally targeting civilians for the third time in a decade.

The reason why the first sentence is difficult to comprehend whereas the second sentence is easy is because the first involves a large number of nested head-final constructions.

One of the reasons why Sanskrit literature is difficult to translate is excessive usage of head finals especially through compound words to create vivid effects. For example, SANSKRIT PHRASE:अज्ञाप-रोग​-पीडित​-धनिकः LITERAL TRANSLATION: Unknown-disease-stricken-richman.

On reading the first word, you don’t get anything. What is unknown here? On reading the second word, you think the phrase is about an unknown disease. But upon seeing the next word, you realise that it talks about pain and suffering and only the last word breaks the suspense that it is about a rich man. Now that you know that the subject is a rich man, you go backwards and figure out that the rich man is suffering due to an unknown disease. But the head-initial English construction “The man stricken by an unknown disease” sounds straightforward and plain. First, you have a solid idea that the subject is a man. Then you know that he is suffering. Then, you get that he is suffering from an unknown disease (a rare instance where head final construction is followed in English where the noun comes later after the adjective that modifies it).

This principle in linguistics where head-initial constructions are easier for cognition is called the “light-before-heavy” principle. This is because, in a head-final construction, you are holding a heavy MODIFIER together for a long time at once, waiting for the single HEAD (lighter one) to finally come. But in a head-initial construction, you are introduced to the short HEAD first and gradually build up the load of its modifiers.

Panini and Light-Before-Heavy

This light-before-heavy principle is also important in mere listing of items. For example, why do we say “bread and butter” not not “butter and bread”? Look at the following lists

• bread and butter (not: butter and bread!)
• Tom, Dick and Harry (not: Harry, Dick and Tom!)
• light and darkness (not: darkness and light)
• Tit for Tat (not: tat for tit)

There are many rules governing the ordering of the constituents in a list. Such lists are common in Sanskrit too and are called by the name of dvandva samāsa. It turns out that in such listings too, the light before heavy rule applies. Items with lower number of syllables appear first!! That is why it is “bread” (one syllabled word) and then butter (two syllabled word)! Words with lesser number of syllables (light) come before ones with higher number of syllables (heavy). What if two words are of equal syllable numbers? Then refer this paper to decide¹. You will then get why we say “Hari-hara” and not “Hara-Hari”! This principle of light-before-heavy was discovered by none other than Panini himself. To quote eminent language psychologist Steven Pinker from his book “The Sense of Style”: ( I highly recommend all books by Steven Pinker - my most favourite author in linguistics)

“Human comprehension demands topic before comment and given before new…The Scottish prayer asks the Lord to deliver us from “ghoulies and ghosties and long-leggedy beasties and things that go bump in the night”—not from “things that go bump in the night and long-leggedy beasties and ghoulies and ghosties.” The order fits with our cognitive processes: it’s taxing to work on a big heavy phrase (things that go bump in the night) while you are holding in memory an incomplete bigger phrase it’s part of (in this case, the four-part coordination embracing things, beasties, ghoulies, and ghosties). A big heavy phrase is easier to handle if it comes at the end, when your work assembling the overarching phrase is done and nothing else is on your mind. (It’s another version of the advice to prefer right-branching trees over left-branching and centre-embedded ones.) Light-before-heavy is one of the oldest principles in linguistics, having been discovered in the fourth century BCE by the Sanskrit grammarian Pāṇini It often guides the intuitions of writers when they have to choose an order for items in a list, as in life, liberty, and the pursuit of happiness; The Wild, The Innocent, and The E Street Shuffle; and Faster than a speeding bullet! More powerful than a locomotive! Able to leap tall buildings in a single bound! Topic, then comment. Given, then new”.

In fact, in linguistics, this is called Panini’s law. https://en.wiktionary.org/wiki/Panini%27s_law

Panini mentions this in Astadhyayi 2.2.34 - अल्पाच्तरम् . S.K.Vasu’s commentary reads “In a द्वन्द्व compound, that word-form which has fewer vowels, is to be placed first. Thus प्लक्ष +न्यग्रीध = प्लक्षन्यग्रोधौ ; and धबखदिरपलाशाः”!

I sincerely hope that you have enjoyed this series of articles and got a spark. What I have provided is just a small drop in the ocean of Paninian grammar and I urge you to explore this vast treasure trove of knowledge.

Footnotes

1. World Order by William Cooper and Haj Ross