Sandhi — literally "joining" in Sanskrit — is the process where two words (or parts of words) come together, and the sounds at the boundary merge or change to form a smoother, combined word.
Think of it like this: when two people shake hands, both hands change position slightly to fit together. Sandhi is exactly that — two word-endings and word-beginnings adjusting to "shake hands" cleanly.
The reverse operation — breaking a combined word back into its original parts — is called Sandhi Vichchhed (संधि विच्छेद). This is what the UP Police exam almost always tests. You are given कवीश्वर and asked: what were the two original words? The answer is कवि + ईश्वर.
Why does it matter for your exam? The Hindi section of UP Police Constable routinely places 3-5 marks on Sandhi. The questions follow a tight pattern: either identify the correct Sandhi Vichchhed of a given word, or identify which type of Sandhi a word exemplifies. Both tasks become mechanical once you know the three categories and their most important rules.
Here is the core framework before we go deep:
Every Sandhi question you will ever see in UP Police fits into one of these three categories. Your job in the exam hall is to quickly classify the word, apply the relevant rule, and move on.
When two vowels meet at a word boundary, one of five sub-rules applies. The exam tests these most frequently:
Two identical or near-identical vowels merge into a single long vowel.
Rule: अ/आ + अ/आ = आ | इ/ई + इ/ई = ई | उ/ऊ + उ/ऊ = ऊ
Examples:
राम + अयन = रामायण (अ + अ = आ)कवि + ईश्वर = कवीश्वर (इ + ई = ई) — this was a 2024 PYQभानु + उदय = भानूदय (उ + उ = ऊ)अ/आ meets इ/ई, उ/ऊ, or ऋ — the vowel "upgrades" to its Guna equivalent.
Rule: अ/आ + इ/ई = ए | अ/आ + उ/ऊ = ओ | अ/आ + ऋ = अर्
Examples:
देव + इंद्र = देवेंद्रमहा + उत्सव = महोत्सवमहा + ऋषि = महर्षिअ/आ meets ए/ऐ or ओ/औ — the vowel upgrades further.
Rule: अ/आ + ए/ऐ = ऐ | अ/आ + ओ/औ = औ
Examples:
एक + एक = एकैकमहा + औषध = महौषधइ/ई, उ/ऊ, or ऋ meets a different vowel — the first vowel converts to its semi-vowel.
Rule: इ/ई → य | उ/ऊ → व | ऋ → र
Examples:
अति + अधिक = अत्यधिक (इ → य)अनु + अय = अन्वय (उ → व)ए, ऐ, ओ, औ meets any vowel — the diphthong splits.
Rule: ए → अय | ऐ → आय | ओ → अव | औ → आव
The key example for your exam: पो + अन = पवन (ओ → अव). This appeared as a 2024 PYQ — पवन is Svar Sandhi via Ayan.
Look — Vyanjan Sandhi rules are numerous, but UP Police tests only a handful of patterns. Focus on these:
त्/द् Assimilation RuleWhen त् or द् meets a consonant of a different class, it changes to match.
त् + ल = ल्ल → उत् + लास = उल्लास (2019 PYQ)त् + श = च्छ → उत् + श्वास = उच्छ्वाससत् + जन = सज्जनA stop consonant at the end of the first word converts to the nasal of its own वर्ग (class) when a nasal begins the next word, or when conditions require nasalization.
वाक् + मय = वाङ्मयCertain prefixes ending in consonants cause the next letter to double.
उत् + लास = उल्लास — here त् drops and ल doubles.This is the highest-frequency category in recent UP Police PYQs. The Visarga (ः) at the end of a word behaves differently depending on what follows it.
Key rules to memorize:
| Situation | Result | Example |
|---|---|---|
| Visarg + च/छ | Visarg → श् | निः + चल = निश्चल |
| Visarg + क/प | Visarg stays or → ष् | निः + कपट = निष्कपट |
| Visarg + त/थ | Visarg → स् | निः + तेज = निस्तेज |
| Visarg after अ + voiced consonant or vowel | Visarg → ओ | मनः + बल = मनोबल |
| Visarg after अ + voiced consonant or vowel | Visarg → ओ | वय: + वृद्ध = वयोवृद्ध |
| Visarg after अ + अ | Visarg drops, second अ becomes ओ | यशः + अभिलाषा = यशोभिलाषा |
The pattern to fix in your head: अः before a voiced sound (swar or voiced vyanjan) → ओ. This gives you मनोबल, वयोवृद्ध, मनोहर, यशोदा — all Visarg Sandhi.
For अन्तर्निहित: the source is अन्तः + निहित. The Visarga before न (a voiced nasal) converts — this is Visarg Sandhi. Not Vyanjan, not Svar.
Whenever you see a compound word ending in ...o... (ओ sound in the middle), ask: could the first part originally end in ः? If yes, it is Visarg Sandhi. This catches मनोबल (मनः), वयोवृद्ध (वयः), मनोहर (मनः), यशोगान (यशः) in under 5 seconds. Standard analysis: 30 seconds of rule-checking. Pattern recognition: 5 seconds.
When two vowels from the same family meet (both इ-family or both उ-family or both अ-family), the result is simply the long version of that vowel. You do not need to think — if कवि + ईश्वर, both are इ-family → result has ई in the middle → कवीश्वर. Eliminates all 3 wrong options in about 8 seconds versus 25 seconds of applying the full Dirgha rule from scratch.
पवन, भवन, नयन — whenever you see a word with व or य sitting between two vowel-sounds and the word looks like it could come from a verb root, check for Ayan Sandhi. पो + अन = पवन (ओ → अव). The examiner loves this because candidates confuse it with Vyanjan Sandhi. Eliminating the three non-Svar options takes 10 seconds once you recognize the ओ → अव fingerprint. Standard option-by-option checking: 40 seconds.
In Sandhi Vichchhed questions involving Visarg words, the correct option will always show ः (the visarga colon) in the first part: अन्तः + निहित, मनः + बल, वय: + वृद्ध. Any option that writes अन्त + निहित (without ः) is automatically wrong. Scan options for ः first — this eliminates 2-3 options in 5 seconds without any rule application. Standard method: reading and evaluating all 4 options = 35 seconds.
The prefix उत् is one of the most tested Vyanjan Sandhi sources. When you see words like उल्लास, उच्चारण, उज्जवल, उत्सव — break the first syllable. उल् came from उत् + ल, उच् from उत् + च, उज् from उत् + ज. Recognizing this prefix pattern handles a whole cluster of Vyanjan Sandhi questions. Step count: 2 steps (identify उत् root → write vichchhed) versus 5 steps (apply full consonant class rules). Saves approximately 20 seconds per question.
When a Sandhi question appears in the exam hall, run this decision tree:
Step 1 — Is this a Vichchhed question or a "which type" question?
Step 2 — Vichchhed shortcut path:
ओ in the middle? → Suspect Visarg Sandhi → look for ः option.ल्ल, ज्ज)? → Vyanjan Sandhi → look for उत् or similar prefix.Step 3 — "Which type" shortcut path:
Scan the options for ः first before doing any rule analysis. This alone eliminates wrong choices in Visarg questions.
Total time target: 45-60 seconds per Sandhi question.
Why this question: The most direct test of Dirgha Sandhi — two इ-family vowels merging. Appears in 2024 paper. Understanding this removes confusion about options that offer ई in the wrong position.
Solving path: कवीश्वर — see the long ई in the middle. Split: what comes before ई-sound? कवि (short इ). What comes after? ईश्वर (starts with long ई). इ + ई = ई (Dirgha Sandhi). Option D: कवि + ईश्वर. Eliminate options that show कवी in the first part — the original word had short इ, which is the whole point of Dirgha Sandhi changing it.
Why this question: Classic Visarg Sandhi vichchhed. Candidates lose marks by writing अन्त + निहित (missing the ः). The ः is non-negotiable in the vichchhed.
Solving path: अन्तर्निहित — the र् in the middle is a giveaway. In many Visarg Sandhi cases, ः before a voiced consonant gives र् (e.g., अन्तः + निहित = अन्तर्निहित). Scan options for ः → only Option C has अन्तः + निहित. Done. 10 seconds.
Why this question: Tests Visarg Sandhi type-identification. Candidates confuse मनोबल with Dirgha or Guna Sandhi because of the ओ sound.
Solving path: मनोबल = मनः + बल. The first part ends in ः (visarga). Before voiced consonant ब, the visarga of मनः (where the vowel before ः is अ) converts to ओ. This is definitionally Visarg Sandhi. Answer: Option B.
Why this question: Tests Ayan Sandhi identification — the trickiest Svar Sandhi subtype. पवन looks like a simple word, not obviously a sandhi product.
Solving path: Check each option. संयोग — सम् + योग, the म् changes — Vyanjan Sandhi. नमस्कार — नमः + कार, has ः — Visarg Sandhi. मनोहर — मनः + हर, has ः — Visarg Sandhi. पवन — पो + अन, where ओ → अव (Ayan Sandhi, a subtype of Svar Sandhi). Answer: Option D.
Why this question: Tests Visarg Sandhi vichchhed for वयोवृद्ध. The trap is writing वयो: or वाया: in the first part — both wrong.
Solving path: वयोवृद्ध — see ओ in the middle. Apply the Visarg alarm: original first word likely ended in ः. The original word is वयस् → in sandhi context written as वय:. Before व (voiced), अः → ओ. So वय: + वृद्ध = वयोवृद्ध. Answer: Option B वय: + वृद्ध. Eliminate वयो: — that would mean the original already had ओ, which is not the source form.
Why this question: Vyanjan Sandhi vichchhed for उल्लास — tests the उत् prefix pattern and त् + ल = ल्ल rule.
Solving path: उल्लास — doubled ल्ल is the signal. Which prefix produces doubling before ल? उत् — because त् + ल → ल्ल. So vichchhed = उत् + लास. Answer: Option D. Eliminate उल + लास — that would require उल to be a prefix, which it is not.
Writing the long form in the first part of vichchhed. For कवीश्वर, writing कवी + ईश्वर instead of कवि + ईश्वर. The whole point of Dirgha Sandhi is that two shorter vowels merged — the original word had the shorter form. Always restore the shorter vowel in the first part.
Missing the Visarga (ः) in Visarg Sandhi vichchhed. Writing मन + बल or अन्त + निहित instead of मनः + बल or अन्तः + निहित. The ः must appear in the answer — scan for it in the options first.
Confusing मनोहर, मनोबल type words with Svar Sandhi. These have ओ in the middle but they are Visarg Sandhi, not Guna Sandhi. The test: does the first part end in ः in the vichchhed? If yes, Visarg Sandhi — period.
Misidentifying पवन, भवन, नयन as Vyanjan Sandhi. The व and य in these words look like consonants, but they arose from vowel-to-semivowel conversion (Ayan Sandhi). These are Svar Sandhi examples, not Vyanjan.
Confusing उल्लास vichchhed as उल + लास. The prefix उल does not exist in Sanskrit-origin Hindi prefixes. The correct first part is always उत् — which then triggers the consonant shift rule. When in doubt, think: which real prefix could produce this?
Treating all ओ-containing words as Visarg Sandhi. गोपाल, भोजन — these have ओ but are not sandhi products at all, or they follow different rules. The Visarg alarm works only when you can find a source word ending in ः. If no such word exists, look for Guna or other Svar rules.