This is some selected parts of “Fact 14 – Sanskrit alphabet is scientifically arranged” excerpt from my upcoming book, “Our Mother Tongue: 108 Facts about Sanskrit”.
Our mother tongue (Sanskrit) has around forty consonant sounds and around 9 vowel sounds (14 if you consider long and short vowels separately). If we add tones or pitch (as in the Vedas), since each vowel can have three different tones, the total number of sounds rises to around eighty. Of course, these sounds can be combined in various ways to give many (hundreds of) combination sounds. The complete list of sounds is given below in the table. [I have used an expanded version (to accommodate alveolar sounds etc.) of the table format that is used by Bucknell.]
Vowel sounds are made with the mouth open, allowing for fairly free flow of air through it, while sounding consonants involves constricting the airflow in different locations of the mouth. For example, when you sound a vowel like a, you can feel your mouth letting the air from the lungs and larynx flow freely with the mouth open. When you sound vowels like i, though you let the air flow freely, you shape the flow of air though a passage between the tongue and the palate, while when you sound a u, you let the air through shaped lips.
However, when you sound a consonant like say b, you can feel your lips briefly stopping the flow of air and then letting go.
Unlike other languages, Sanskrit has a unique, vocalisation-based classification of its sounds (and alphabet). Other Indian languages follow the same approach. The alphabet classification representation in the picture in one which takes into account sound articulation positions.
Look at the topmost row labelled “open”. The a-vowels (short a and long ā) are vowels sounded with an open throat with no obstruction whatsoever. Hence, they are in a separate row with a couple of other open sounds, which we will see later.
Let us now take the row labelled guttural or velar. These sounds are made by stopping the airflow momentarily with a constriction produced by moving the back part of the tongue against the soft palate (or velum). There are six columns in this row. Look at k and g in this row. [Note that in the Devanāgarī script, a consonant is represented with an a sounded at the end. A sound without the a sound needs the virāma tacked on. So, क ka; but क् k] The main difference between these two, k and g, is that k is unvoiced (voiceless), called a surd, while g is voiced, called a sonant. The main difference between a surd and a sonant is that while you sound a sonant, your vocal cords vibrate, while when you sound a surd, they do not. Now let us compare two other columns in the row. Take k and kh. kh is aspirated when you sound it. It means that a strong breath of air accompanies the sounding of the consonant. It is almost as if the k is joined with a h.
The column labelled “nasal” in this row, has the consonant ṅ that sounds like the ng in the English word “sing”. This sound is called a nasal because the air that is briefly stopped at the velum, is diverted through the nose.
So, in the row labelled “velar” or “guttural”, there are five kinds of sounds. An unaspirated surd, an aspirated surd, an unaspirated sonant, an aspirated sonant and a nasal.
There are four more main rows in the Sanskrit classification of consonants. The palatal, where the flow of air is obstructed by the middle of the tongue against the hard palate; the lingual or the retroflex, where the obstruction is caused by the tongue (slightly curled) touching against the part of the mouth between the alveolar ridge and the hard palate; the dental, produced by an obstruction caused by the front part of the tongue against the upper teeth; and the labial, produced by closing the two lips.
Corresponding to each row, there are five kinds of sounds, as seen in the velar row.
One extra row shown in this table, that is not normally shown in Sanskrit tables, is a row for alveolar consonants. This has the n that is normally pronounced in the inner and ending positions in a word. The n in the beginning is normally pronounced as a dental sound. You can hear this difference when in Hindi, you say the n in “namaskar” as against “sapna”. In Tamil, these two sounds have separate alphabetic characters – ந and ன respectively.
Note that Indian languages (with the exception of Malayalam) do not have the alveolar t, the way t is pronounced in English. Indians normally use the retroflex ṭ when they want to say t. The word “tea” is pronounced “ṭea” by most Indians.
In addition to these five columns, there are two more columns of consonants. The ones in the column called fricatives are sounds produced when air is forced through small gaps in the mouth. The sibilants are fricatives that are produced when air is forced through a small gap between the upper and lower teeth. ś, ṣ and s are sibilants. Depending on where these are articulated they are placed in the appropriate row. For example, ś is a palatal, ṣ is a retroflex and s is a dental. The fricatives are not voiced.
The i-vowels are classified with the palatals since the airflow is shaped through the palate. These vowels shade into the palatal consonants through their semivowel y.
The u-vowels are similarly related to the labial class. Their semivowel is v. [Note that the semivowel v is sometimes classified as a voiced labiodental fricative, like in English.]
The r̥ and the l̥ vowels represent the vocalic r and l, that is, these are the vowels corresponding to the semivowels r and l and are therefore placed in the lingual and dental classes respectively.
You should note that the a-vowels do not have a corresponding semivowel, which the other two vowels have.
The so-called diphthongs e and o are made by incrementing i and u respectively with an a. Note that these are long sounds. In Sanskrit, they lack the corresponding short e and o that is seen in Tamil and other Dravidian languages. The corresponding higher increments of i and u are ai and au. They are classified along with the corresponding vowels.
e and o are called the guṇa of the i- and u-vowels respectively, and ai and au are called the vr̥ddhi respectively. The r̥- and the l̥-vowels also have guṇa and vr̥ddhi. The guṇa and vr̥ddhi of the r̥-vowels are ar and ār (and sometimes ra) respectively and those of the l̥-vowels are al and āl (and sometimes la) respectively. We will learn about the use and need for the guṇa and vr̥ddhi in later chapters.
The semivowels function both as vowels and as consonants. They are articulated with relatively less, and relatively more obstruction than the consonants and vowels respectively. They are also slotted into their corresponding rows based on their point of articulation. All semivowels, like the vowels, are voiced.
Another sound is the breathing or aspiration h, which, though it is classified among the gutturals, is normally sounded at the position of the following vowel. It is the aspirate part of the aspirated consonants. In Sanskrit, it has more the characteristic of a sonant than a surd. [In English it is voiceless.]
There are four sounds that I have not talked about so far. They are shown shaded in the alphabet table. On the top left corner is the visarga. The visarga is not an independent sound. It is a euphonic (Sandhi) change of final s and r. [It may also take the place of s in some internal combinations]
The visarga is almost like a voiceless h, articulated at the point of the previous vowel (not the following vowel as in the case of h). [I remember our family purohita (teacher) telling us “break at the visarga!” It is almost like the glottal stop of Cockney English in the word “better” pronounced, “be’er”. In fact, small children sometimes use this glottal stop instead of “s”. My cousin’s kid used to say that she studied in P-eḥ-B-B (PSBB) school.] Note that the vowel should not shadow after the visarga. [Some traditions in India actually pronounce the visarga as an h with the vowel before the visarga repeated after it. For example, rāmaḥ is pronounced as rāmaha]. The visarga is also called the visarjanīya (the Prātiśākhyas and Pāṇini normally use this term)
The two others in the fricatives column are the ẖ and the ḫ, called the jihvāmūlīya and the upadhmānīya respectively. These are how the visargas are pronounced when before a guttural sound (like k), and when before a labial sound (like p) respectively. The ẖ is pronounced like the ch in the Scottish word “loch”. And the ḫ, like an English f. Thus, namaḥ ka… is pronounced namaẖ ka… (namachka…) and namaḥ pa… is pronounced namaḫ pa… (namafpa…).
The other sound is the shaded cell in the nasal column. It is called an anusvāra in Sanskrit. It nasalises the preceding vowel. It almost takes the place of a consonant. It occurs at the end of words and in the middle of a word before the sibilants and the h. In actual usage, it is almost sounded like an m or an n. There is another representation of the nasal: m̐. It is represented by a chandrabindu on top. [Some grammarians talk of some intricate differences between the two, but for all practical purposes they are the same.]
The short marker with two Sanskrit letters like cay, khay, ac, ak etc. are Paninean schemes for identifying groups of sounds. We will look at these later.
Note that the vowels have three representations in the Latin script. The first is the unaccented one, the second is the same vowel with the acute accent (the udātta), and the third, the same vowel with the grave accent (the independent svarita). These accents are used only in Vedic Sanskrit. We will look at Vedic accents later.
You can see that the sounds of our mother tongue are arranged in a very scientific manner, based on the point of articulation and aspiration, and on whether it is voiced or not.
Additional Material (not from the book)
Use of Vowels